{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":120,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":120,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"3a9fca90f174","filters":{"venue":"IEEE Transactions on Knowledge and Data Engineering"}},"results":[{"id":"W2096451472","doi":"10.1109/tkde.2005.50","title":"Using AUC and accuracy in evaluating learning algorithms","year":2005,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":2141,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Machine learning; Computer science; Artificial intelligence; Naive Bayes classifier; Decision tree; Measure (data warehouse); Receiver operating characteristic; Algorithm; Bayes' theorem; Data mining; Support vector machine; Bayesian probability","authors":[{"name":"Jin Huang","is_ca":true},{"name":"Charles X. Ling","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09066818198011457,"gpt":0.3642777199882067,"spread":0.2736095380080922,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05874668,0.002383707,0.003281337,0.02158175,0.001275367,0.009896017,0.002416904,0.005252562,0.001491223],"category_scores_gemma":[0.333273,0.0005635749,0.00174501,0.01611619,0.004263908,0.01232692,0.003655167,0.003964902,0.001279264],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002244469,"about_ca_system_score_gemma":0.00238119,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001989118,"about_ca_topic_score_gemma":0.001204542,"domain_scores_codex":[0.8771406,0.06951564,0.01190552,0.006517403,0.03319011,0.001730851],"domain_scores_gemma":[0.6652827,0.2669715,0.01955133,0.01746085,0.02794586,0.002787818],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001333094,0.0003707398,0.1185728,0.002068989,0.002124686,0.0004614014,0.001174812,0.1493007,0.00280197,0.1044201,0.01664843,0.6007223],"study_design_scores_gemma":[0.0002166663,0.002354494,0.05139852,0.001703061,0.0008737101,0.002268721,0.001131045,0.5649844,0.01107189,0.3191436,0.04424226,0.0006117574],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09970058,0.03269701,0.8313965,0.00467087,0.002160357,0.0005362559,0.00195229,0.00288674,0.02399947],"genre_scores_gemma":[0.6527705,0.006437828,0.331517,0.001423285,0.002100779,0.0007827391,0.002346695,0.0007698179,0.001851366],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.05874668,"threshold_uncertainty_score":0.3106859,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2151953639","doi":"10.1109/tkde.2005.166","title":"Fast algorithms for frequent itemset mining using FP-trees","year":2005,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":552,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Data mining; Traverse; Association rule learning; Algorithm; Tree (set theory); Data structure; Tree structure; Trie; Prefix; Binary tree; Mathematics","authors":[{"name":"Gösta Grahne","is_ca":true},{"name":"Jun Zhu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05517782314487753,"gpt":0.3067297556925058,"spread":0.2515519325476283,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006331121,0.002710182,0.003150566,0.007442361,0.002156128,0.002559494,0.003521581,0.002636136,0.004388689],"category_scores_gemma":[0.02853196,0.001748009,0.002424892,0.01038756,0.000772853,0.006881258,0.001551706,0.002630675,0.005217967],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001013653,"about_ca_system_score_gemma":0.002183775,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002423296,"about_ca_topic_score_gemma":0.002118937,"domain_scores_codex":[0.9952855,0.001128476,0.0007426538,0.0007659665,0.001780671,0.000296862],"domain_scores_gemma":[0.98015,0.0130163,0.001323591,0.001565065,0.003760589,0.0001844281],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004819868,0.0002638352,0.003242886,0.0008821056,0.0003762087,0.0003620489,0.0003072555,0.0569064,0.004409283,0.01765049,0.01712701,0.8979905],"study_design_scores_gemma":[0.000334072,0.0003799578,0.001654827,0.00027508,0.0002647925,0.001877789,0.0002277501,0.8398507,0.01417568,0.1123861,0.02840991,0.0001634568],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004194064,0.001256618,0.990008,0.0002194829,0.0001194287,0.000227195,0.0004799875,0.002760461,0.0007347819],"genre_scores_gemma":[0.02491221,0.0008922489,0.9714082,0.00008709006,0.0001118427,0.00055577,0.001323147,0.0001381519,0.0005714179],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007442361,"threshold_uncertainty_score":0.03348255,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2170595610","doi":"10.1109/tkde.2004.1269594","title":"CAIM discretization algorithm","year":2004,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Rough Sets and Fuzzy Logic","field":"Computer Science","cited_by":449,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Discretization; Discretization of continuous features; Algorithm; Computer science; Decision tree; Class (philosophy); Tree (set theory); ID3 algorithm; Decision tree learning; Artificial intelligence; Incremental decision tree; Mathematics; Discretization error","authors":[{"name":"Lukasz Kurgan","is_ca":true},{"name":"Krzysztof J. Cios","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01997980423624899,"gpt":0.2493373957629373,"spread":0.2293575915266883,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00176727,0.0008299684,0.001390545,0.002071298,0.00109106,0.002850392,0.002671128,0.001993194,0.01429722],"category_scores_gemma":[0.008092287,0.0005263722,0.001575244,0.003063163,0.00071763,0.002193318,0.002139706,0.002620577,0.005243058],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001344577,"about_ca_system_score_gemma":0.002009822,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003637834,"about_ca_topic_score_gemma":0.00307469,"domain_scores_codex":[0.9979939,0.0003345311,0.0002123694,0.0004442981,0.0008521936,0.0001626923],"domain_scores_gemma":[0.9974064,0.00090848,0.0001262579,0.0005806052,0.0009068397,0.00007136902],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002864814,0.0001062742,0.001607387,0.0004224239,0.0001046191,0.0001889156,0.0002417657,0.1309944,0.004943051,0.08943529,0.0501657,0.7215037],"study_design_scores_gemma":[0.0000851648,0.00007579455,0.00051931,0.0000864258,0.00003691103,0.000349854,0.00009625968,0.847951,0.006225958,0.06191643,0.08261566,0.00004110195],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002129558,0.0004733619,0.9866919,0.0002338068,0.0002156845,0.0001487855,0.0004903934,0.001423748,0.008192882],"genre_scores_gemma":[0.04766588,0.0004711386,0.9414204,0.0003100239,0.0001236914,0.0004409626,0.002082885,0.0002904763,0.007194489],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01429722,"threshold_uncertainty_score":0.04782897,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2137763598","doi":"10.1109/tkde.2004.58","title":"Efficient phrase-based document indexing for Web document clustering","year":2004,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":320,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Document clustering; Computer science; Cluster analysis; Vector space model; Search engine indexing; Phrase; Information retrieval; Document classification; Data mining; tf–idf; Artificial intelligence; Term (time)","authors":[{"name":"Khaled M. Hammouda","is_ca":true},{"name":"Mohamed S. Kamel","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02028388213544066,"gpt":0.2679597706680517,"spread":0.247675888532611,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001256536,0.000784139,0.001399069,0.004405822,0.001098526,0.00160535,0.001782839,0.001007364,0.002336452],"category_scores_gemma":[0.007110163,0.0004591109,0.001024594,0.009499787,0.0006611206,0.002796623,0.001643425,0.001270631,0.002618118],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001103874,"about_ca_system_score_gemma":0.001569829,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004000303,"about_ca_topic_score_gemma":0.004178685,"domain_scores_codex":[0.9983574,0.0004335535,0.0001242489,0.0002648279,0.0007371228,0.00008283414],"domain_scores_gemma":[0.9980203,0.0006195031,0.0001676637,0.0005740468,0.0005630313,0.00005553494],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002013628,0.0001207917,0.0008992539,0.0004222789,0.0001033649,0.0001075881,0.0002800754,0.0887168,0.02255181,0.05916213,0.01699824,0.8104363],"study_design_scores_gemma":[0.00004592617,0.00008310653,0.0005826834,0.00002482439,0.0000490034,0.0002336378,0.00007138492,0.9032925,0.008853515,0.07417627,0.01253762,0.00004959584],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003313171,0.0003755327,0.993883,0.0000627565,0.00003433012,0.00009091949,0.0003639852,0.00116452,0.0007118771],"genre_scores_gemma":[0.03637255,0.0005158997,0.9601054,0.00004729038,0.00007315654,0.0002402213,0.001676647,0.00021147,0.0007574298],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004405822,"threshold_uncertainty_score":0.008009255,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2137525015","doi":"10.1109/tkde.2006.80","title":"Pattern discovery of fuzzy time series for financial prediction","year":2006,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Stock Market Forecasting Methods","field":"Decision Sciences","cited_by":179,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Research Council Canada; University of Illinois at Urbana-Champaign; National Science Council","keywords":"Fuzzy logic; Investment (military); Computer science; Fuzzy set; Time series; Data mining; Finance; Process (computing); Representation (politics); Artificial intelligence; Machine learning; Economics","authors":[{"name":"Chiung-Hon Leon Lee","is_ca":false},{"name":"Alan Liu","is_ca":false},{"name":"Wen-Sung Chen","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05021834214437013,"gpt":0.3252064956071853,"spread":0.2749881534628152,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001185328,0.0004194577,0.0008214544,0.002315829,0.0003820669,0.0008857774,0.0007849008,0.0006189582,0.001803184],"category_scores_gemma":[0.006341463,0.0002097847,0.0007005946,0.00291614,0.0003780539,0.001269065,0.000471423,0.0006573962,0.0004150425],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005259755,"about_ca_system_score_gemma":0.0006395906,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003966072,"about_ca_topic_score_gemma":0.002775914,"domain_scores_codex":[0.9992734,0.0001749341,0.00008597284,0.0001482859,0.0002779004,0.00003949035],"domain_scores_gemma":[0.9982837,0.0009505914,0.0001503113,0.0002180497,0.000358818,0.00003854085],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002601297,0.0001522169,0.00584677,0.0003771207,0.0001326603,0.0004251004,0.0002878022,0.08823533,0.01373808,0.02588424,0.004450014,0.8602105],"study_design_scores_gemma":[0.00001601718,0.00004758636,0.001744858,0.00002712511,0.00003041887,0.0001476896,0.0000562565,0.9713582,0.004761319,0.01839968,0.003388354,0.00002238254],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02834515,0.0007836768,0.9674225,0.0003319578,0.00007133339,0.00009625843,0.0004018303,0.001055899,0.001491355],"genre_scores_gemma":[0.358053,0.0009426634,0.6382083,0.00008908498,0.0000666227,0.0001849191,0.0009564295,0.00005057524,0.001448395],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003966072,"threshold_uncertainty_score":0.007885933,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2142804492","doi":"10.1109/tkde.2011.138","title":"A Lightweight Algorithm for Message Type Extraction in System Application Logs","year":2011,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Software System Performance and Reliability","field":"Computer Science","cited_by":146,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Message passing; Partition (number theory); Event (particle physics); Algorithm; Data type; Web log analysis software; Task (project management); Data mining; Theoretical computer science; Distributed computing; Programming language; Operating system; The Internet","authors":[{"name":"Adetokunbo Makanju","is_ca":true},{"name":"A. Nur Zincir‐Heywood","is_ca":true},{"name":"Evangelos Milios","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02653254209557903,"gpt":0.2630677020256542,"spread":0.2365351599300751,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001830768,0.001515722,0.001034823,0.004903258,0.001186591,0.002134086,0.002209128,0.001232174,0.00407003],"category_scores_gemma":[0.009325748,0.0007872199,0.001218916,0.00322907,0.000653829,0.003696904,0.002107933,0.001653233,0.004285804],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001249314,"about_ca_system_score_gemma":0.002465335,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002370723,"about_ca_topic_score_gemma":0.003575035,"domain_scores_codex":[0.9976717,0.0002946183,0.0003437117,0.0005074628,0.001034772,0.0001476574],"domain_scores_gemma":[0.9948461,0.002068999,0.0007246453,0.001211239,0.001026365,0.0001227628],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000387824,0.0002138888,0.003583157,0.0004180885,0.00007312309,0.0001564059,0.0002711779,0.01086109,0.02097171,0.00651547,0.01655197,0.9399962],"study_design_scores_gemma":[0.0001942763,0.0002438284,0.003424596,0.0001157402,0.00007666229,0.0007593101,0.0002343091,0.8646445,0.05278838,0.03601812,0.04138619,0.0001141069],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00457055,0.0001303327,0.978224,0.000104454,0.00004499501,0.0003118023,0.0006708455,0.01541815,0.0005249332],"genre_scores_gemma":[0.03595946,0.00009045711,0.9579713,0.00009190832,0.00004448365,0.0004797845,0.002799107,0.0006694241,0.001894002],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004903258,"threshold_uncertainty_score":0.01361567,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2157751754","doi":"10.1109/tkde.2006.131","title":"Test strategies for cost-sensitive decision trees","year":2006,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":145,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Hong Kong University of Science and Technology","keywords":"Computer science; Medical diagnosis; Decision tree; Machine learning; Test (biology); Artificial intelligence; Test case; Process (computing); Medical costs; Data mining; Health care; Medicine","authors":[{"name":"Charles X. Ling","is_ca":true},{"name":"Victor S. Sheng","is_ca":true},{"name":"Qiang Yang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02750885075641095,"gpt":0.2859938931044155,"spread":0.2584850423480046,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007781953,0.001092864,0.001188462,0.001620086,0.0005341926,0.001189397,0.002172521,0.001730973,0.001797442],"category_scores_gemma":[0.03599237,0.0004882198,0.000577639,0.00102549,0.001073803,0.002969886,0.001227765,0.001445732,0.0002842043],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00130321,"about_ca_system_score_gemma":0.001230924,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00148555,"about_ca_topic_score_gemma":0.0013073,"domain_scores_codex":[0.9949924,0.003112132,0.000310227,0.0004144892,0.0009444717,0.0002262974],"domain_scores_gemma":[0.9725062,0.02344914,0.0009926545,0.000857797,0.001812992,0.0003813519],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0005287331,0.0003975281,0.005849201,0.0002089638,0.0001331705,0.000288524,0.0002483365,0.6964708,0.002382136,0.05256984,0.002631424,0.2382913],"study_design_scores_gemma":[0.0000548413,0.000108452,0.0002303372,0.00001690224,0.00002067982,0.00004571373,0.00001977734,0.9624189,0.001105007,0.03561632,0.0003523176,0.0000106817],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06898145,0.0006586039,0.9269359,0.0007223111,0.00004199526,0.0003589235,0.0001284137,0.0006175368,0.001554811],"genre_scores_gemma":[0.7107391,0.0001996045,0.2871736,0.0003800477,0.00005016589,0.000427791,0.0002686971,0.00008477883,0.0006762366],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007781953,"threshold_uncertainty_score":0.0411554,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2129546202","doi":"10.1109/tkde.2009.174","title":"An Efficient Concept-Based Mining Model for Enhancing Text Clustering","year":2009,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":131,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Sentence; Natural language processing; Phrase; Term (time); Similarity (geometry); Semantics (computer science); Cluster analysis; Artificial intelligence; Document clustering; Meaning (existential); Word (group theory); Measure (data warehouse); Information retrieval; Similarity measure; Semantic similarity; Data mining; Linguistics","authors":[{"name":"Shady Shehata","is_ca":true},{"name":"Fakhri Karray","is_ca":true},{"name":"Mohamed S. Kamel","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02779020595604768,"gpt":0.2810140212831417,"spread":0.253223815327094,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002401253,0.00120815,0.00156362,0.002691224,0.001174269,0.001540669,0.003165118,0.001481713,0.001851873],"category_scores_gemma":[0.005989402,0.000546024,0.0016378,0.003836383,0.0006500463,0.003774294,0.001281214,0.001601557,0.001382491],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00146753,"about_ca_system_score_gemma":0.002304716,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006205599,"about_ca_topic_score_gemma":0.006492814,"domain_scores_codex":[0.998032,0.0004218387,0.0001275843,0.0004775755,0.0008587477,0.00008221613],"domain_scores_gemma":[0.9980314,0.0007630727,0.0001266179,0.0001815688,0.0008494802,0.00004779686],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002577812,0.0004638395,0.002256577,0.0004817782,0.0002271726,0.0003603792,0.0006281056,0.3258717,0.01287088,0.05980281,0.01117904,0.5855999],"study_design_scores_gemma":[0.00001140514,0.00002872181,0.0001354591,0.00001142023,0.00001541556,0.000096718,0.00002659438,0.9864913,0.001324792,0.009396266,0.002448864,0.00001302773],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004177923,0.0001679203,0.9939487,0.00014104,0.00003156213,0.0001640004,0.0001361474,0.0005482856,0.0006844412],"genre_scores_gemma":[0.06240765,0.0003027297,0.9339467,0.0001623432,0.00004591308,0.0005506253,0.0007043881,0.00008416014,0.001795447],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006205599,"threshold_uncertainty_score":0.01269919,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1969381345","doi":"10.1109/tkde.2014.2320725","title":"Malware Propagation in Large-Scale Networks","year":2014,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Network Security and Intrusion Detection","field":"Computer Science","cited_by":123,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"National Natural Science Foundation of China","keywords":"Malware; Computer science; Pareto distribution; Scale (ratio); Scale-free network; Network security; Exponential function; Computer security; Exponential distribution; Complex network; Theoretical computer science; Data mining; Mathematics; Statistics","authors":[{"name":"Shui Yu","is_ca":false},{"name":"Guofei Gu","is_ca":false},{"name":"Ahmed Barnawi","is_ca":false},{"name":"Song Guo","is_ca":false},{"name":"Ivan Stojmenović","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01172611878137981,"gpt":0.2281549863714376,"spread":0.2164288675900578,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001258859,0.000634666,0.0005657565,0.001147521,0.0007546861,0.001063146,0.0006734036,0.001089033,0.0005554109],"category_scores_gemma":[0.00577708,0.0003974769,0.000556535,0.0005113003,0.00121301,0.002692821,0.001041825,0.0008345139,0.000123188],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001234123,"about_ca_system_score_gemma":0.0005139449,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005125638,"about_ca_topic_score_gemma":0.003729693,"domain_scores_codex":[0.9993713,0.0002220395,0.00002298244,0.0001172668,0.0001768832,0.00008941985],"domain_scores_gemma":[0.9960627,0.002691215,0.0006094266,0.0001871715,0.0003482098,0.0001013088],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00005366512,0.00005928472,0.007736172,0.00007200299,0.00005214076,0.0003879463,0.0002519513,0.9449491,0.005242622,0.02593654,0.0006603274,0.01459816],"study_design_scores_gemma":[0.000003436258,0.00001539976,0.0007024295,0.000005364297,0.000006786775,0.00007022965,0.00005239856,0.9901978,0.0005323788,0.008213243,0.0001941029,0.00000645778],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4289699,0.0009711957,0.5650257,0.0007798115,0.00003611748,0.0001147129,0.0001021652,0.0003266936,0.003673744],"genre_scores_gemma":[0.9847787,0.000544317,0.01319451,0.0000666609,0.00002470868,0.00004066138,0.00005525266,0.00002049008,0.001274749],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005125638,"threshold_uncertainty_score":0.01019162,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2085909071","doi":"10.1109/tkde.2010.226","title":"Privacy Preserving Decision Tree Learning Using Unrealized Data Sets","year":2010,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":112,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Victoria","funders":"","keywords":"Computer science; Decision tree; Incremental decision tree; Tree (set theory); Decision tree learning; Information privacy; Data mining; Artificial intelligence; Machine learning; Computer security; Mathematics","authors":[{"name":"Pui Kuen Fong","is_ca":true},{"name":"Jens H. Weber-Jahnke","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06209211290478007,"gpt":0.3253722159210252,"spread":0.2632801030162451,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009581977,0.0005089937,0.001399883,0.001005814,0.0008776012,0.003297595,0.001985683,0.001070112,0.0008058589],"category_scores_gemma":[0.02729378,0.0005236139,0.001191462,0.00182657,0.002293551,0.007855451,0.003162605,0.002478389,0.0002995815],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001067876,"about_ca_system_score_gemma":0.001432166,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005229284,"about_ca_topic_score_gemma":0.00039373,"domain_scores_codex":[0.9894794,0.004498864,0.001098085,0.001676278,0.002681881,0.000565639],"domain_scores_gemma":[0.9749035,0.01301077,0.00167479,0.008708278,0.001350227,0.0003523534],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001655533,0.0002994697,0.00553314,0.0003063495,0.0002326516,0.0006660348,0.000838251,0.3915553,0.01087179,0.3613155,0.002563084,0.2241629],"study_design_scores_gemma":[0.00004690502,0.0001721119,0.0003860803,0.00002965347,0.00003268141,0.0003737276,0.00009586967,0.7315882,0.01089383,0.2534624,0.002884446,0.00003408494],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03720265,0.0002195256,0.9607078,0.0004440518,0.00003145115,0.00003903318,0.0003020616,0.0002121632,0.0008411471],"genre_scores_gemma":[0.7230715,0.0003789953,0.2737731,0.0002127717,0.00007737637,0.0001469704,0.001020447,0.00005453107,0.00126434],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009581977,"threshold_uncertainty_score":0.05067497,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2142103633","doi":"10.1109/tkde.2006.172","title":"Discovering Frequent Closed Partial Orders from Strings","year":2006,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":93,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China; National Science Foundation","keywords":"Computer science; Multiset; Data mining; String (physics); Pruning; Set (abstract data type); Knowledge extraction; Order (exchange); Theoretical computer science; Mathematics","authors":[{"name":"Jian Pei","is_ca":true},{"name":"Haixun Wang","is_ca":false},{"name":"Jian Liu","is_ca":false},{"name":"Ke Wang","is_ca":true},{"name":"Jianyong Wang","is_ca":false},{"name":"Philip S. Yu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01726091384251803,"gpt":0.2454227689265866,"spread":0.2281618550840686,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001485871,0.0008262871,0.001650773,0.004723179,0.001016624,0.001918172,0.001098614,0.0009625265,0.0009231555],"category_scores_gemma":[0.01501178,0.0005743806,0.001394089,0.004411127,0.0008975301,0.003432033,0.00119239,0.001149728,0.0004451615],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005395763,"about_ca_system_score_gemma":0.001918402,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0015371,"about_ca_topic_score_gemma":0.002619673,"domain_scores_codex":[0.9971616,0.0004672608,0.0004101244,0.0005759661,0.001170885,0.0002142445],"domain_scores_gemma":[0.9869489,0.008455784,0.001401813,0.001316694,0.00153328,0.00034354],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001060183,0.000749558,0.0530154,0.001623452,0.0004457575,0.004391152,0.001666145,0.1408927,0.03525902,0.05267023,0.008462062,0.6997643],"study_design_scores_gemma":[0.00009789277,0.0003136003,0.009357158,0.0001538466,0.0001178326,0.002682037,0.0006911347,0.7034766,0.01903647,0.2536132,0.01036521,0.00009503088],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.272223,0.001294818,0.7156366,0.0006841759,0.00006437194,0.0004384091,0.005819207,0.001985955,0.001853438],"genre_scores_gemma":[0.4113231,0.0007874842,0.5706759,0.0001953525,0.00008160876,0.0004010982,0.01523666,0.0001494501,0.001149406],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004723179,"threshold_uncertainty_score":0.007858098,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2142407957","doi":"10.1109/tkde.2010.152","title":"A Machine Learning Approach for Identifying Disease-Treatment Relations in Short Texts","year":2010,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":88,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Domain (mathematical analysis); Field (mathematics); Set (abstract data type); Health care; Machine learning; Artificial intelligence; Dissemination; Information extraction; Data science","authors":[{"name":"Oana Frunza","is_ca":true},{"name":"Diana Inkpen","is_ca":true},{"name":"Thomas Tran","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03615084399736213,"gpt":0.3036804357149973,"spread":0.2675295917176352,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003703634,0.001449837,0.001076001,0.008528781,0.001250339,0.001702709,0.00126639,0.002082722,0.003239533],"category_scores_gemma":[0.0140339,0.0004173605,0.001540358,0.006561076,0.0008854509,0.003274858,0.001145486,0.001897935,0.002148594],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001003724,"about_ca_system_score_gemma":0.001685072,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002461798,"about_ca_topic_score_gemma":0.003383113,"domain_scores_codex":[0.9955769,0.001649092,0.0007966117,0.001066242,0.0007820696,0.0001291465],"domain_scores_gemma":[0.9887124,0.009198558,0.0008685,0.0004162689,0.0006701357,0.0001340637],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005914831,0.0007850201,0.01008638,0.001691633,0.000497454,0.0008792362,0.0009176021,0.02120553,0.01536142,0.01547331,0.01526602,0.9172449],"study_design_scores_gemma":[0.0002540094,0.0008495509,0.01765995,0.0005058808,0.000660889,0.002134889,0.000790077,0.809005,0.01530815,0.09427962,0.05831947,0.0002324901],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02954315,0.005077667,0.9463759,0.003036985,0.0005575659,0.0009008871,0.007989472,0.002656111,0.003862191],"genre_scores_gemma":[0.2076499,0.00177557,0.7702208,0.0008568141,0.001223494,0.001586039,0.01349694,0.0001221566,0.003068282],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008528781,"threshold_uncertainty_score":0.01958692,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3173856251","doi":"10.1109/tkde.2021.3090664","title":"Hierarchical Multi-View Graph Pooling with Structure Learning","year":2021,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":78,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"National Natural Science Foundation of China","keywords":"Computer science; Pooling; Theoretical computer science; Graph; Upsampling; Artificial intelligence","authors":[{"name":"Zhen Zhang","is_ca":false},{"name":"Jiajun Bu","is_ca":false},{"name":"Martin Ester","is_ca":true},{"name":"Jianfeng Zhang","is_ca":false},{"name":"Zhao Li","is_ca":false},{"name":"Chengwei Yao","is_ca":false},{"name":"Dai Huifen","is_ca":false},{"name":"Zhi Yu","is_ca":false},{"name":"Can Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01904280756439184,"gpt":0.2570408665654081,"spread":0.2379980590010163,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008034958,0.00196224,0.001420159,0.001198712,0.0004647678,0.000834914,0.002450219,0.001201598,0.002380899],"category_scores_gemma":[0.002109559,0.000510866,0.001594871,0.001484156,0.0008658304,0.003501748,0.002005538,0.001548621,0.0007173188],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001078597,"about_ca_system_score_gemma":0.0009897688,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006221428,"about_ca_topic_score_gemma":0.01095699,"domain_scores_codex":[0.9994746,0.00008988207,0.00002375259,0.0002248131,0.0001093919,0.00007764897],"domain_scores_gemma":[0.9994575,0.0001245637,0.00008199184,0.00018872,0.00009560729,0.00005159899],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002470582,0.0002504262,0.001687536,0.0002704109,0.0002936201,0.000228909,0.0001705029,0.3689157,0.02585555,0.02168177,0.01068749,0.569711],"study_design_scores_gemma":[0.00001123942,0.00007947481,0.0002977316,0.000007383136,0.00003334304,0.00004509114,0.00001934375,0.9784133,0.005332912,0.01473794,0.001010508,0.00001178725],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02447323,0.0004601093,0.9706002,0.0001806841,0.00006336943,0.00007065657,0.0002009759,0.002725023,0.001225876],"genre_scores_gemma":[0.5919028,0.0005893334,0.400013,0.0005063817,0.0001176383,0.0001744658,0.00172288,0.0004012823,0.004572208],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006221428,"threshold_uncertainty_score":0.01237041,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2059054797","doi":"10.1109/tkde.2013.88","title":"Mining Statistically Significant Co-location and Segregation Patterns","year":2013,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":73,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Data mining; Correlation; Artificial intelligence; Pattern recognition (psychology); Mathematics","authors":[{"name":"Sajib Barua","is_ca":true},{"name":"Jörg Sander","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02346222270620061,"gpt":0.2662764905445469,"spread":0.2428142678383463,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00239483,0.001139005,0.001899119,0.009241989,0.001363899,0.001828654,0.002109139,0.001680875,0.0017265],"category_scores_gemma":[0.01830677,0.0004978385,0.001421702,0.008087779,0.0008862258,0.002096727,0.001925344,0.001109795,0.001416149],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000629233,"about_ca_system_score_gemma":0.001360489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0029249,"about_ca_topic_score_gemma":0.005168915,"domain_scores_codex":[0.9964766,0.0005696064,0.0004559715,0.001158313,0.001001878,0.0003376984],"domain_scores_gemma":[0.982091,0.01055609,0.002843421,0.001797218,0.002235043,0.0004771958],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001078712,0.0008835203,0.5279152,0.001124015,0.00104656,0.002902363,0.001047757,0.04372813,0.01661319,0.009384167,0.01022223,0.3840542],"study_design_scores_gemma":[0.0001745962,0.0004211813,0.1415484,0.0001971253,0.0005550195,0.004803709,0.002699974,0.7630833,0.02207498,0.05299247,0.0113082,0.0001411225],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.4904813,0.001371556,0.4970025,0.0008132793,0.00009069918,0.0003657822,0.004672589,0.001793343,0.00340911],"genre_scores_gemma":[0.8360413,0.0002656695,0.1568222,0.0001027058,0.00007412263,0.0002427983,0.005346516,0.0000940187,0.001010698],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009241989,"threshold_uncertainty_score":0.01266521,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3185513611","doi":"10.1109/tkde.2022.3220948","title":"Rethinking Graph Auto-Encoder Models for Attributed Graph Clustering","year":2022,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":72,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Cluster analysis; Randomness; Computer science; Adjacency matrix; Correlation clustering; Feature (linguistics); Graph; Clustering coefficient; Encoder; Data mining; Pattern recognition (psychology); Theoretical computer science; Artificial intelligence; Algorithm; Mathematics; Statistics","authors":[{"name":"Nairouz Mrabah","is_ca":true},{"name":"Mohamed Bouguessa","is_ca":true},{"name":"Mohamed Fawzi Touati","is_ca":true},{"name":"Riadh Ksantini","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05141376820259933,"gpt":0.2706461837856772,"spread":0.2192324155830779,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002434805,0.001132671,0.001170299,0.001184415,0.0006724589,0.001636358,0.003568404,0.001916943,0.002486895],"category_scores_gemma":[0.008733813,0.0007747288,0.0012299,0.0012885,0.001603412,0.003667807,0.00271957,0.003051905,0.00137129],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00204119,"about_ca_system_score_gemma":0.002103894,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007679496,"about_ca_topic_score_gemma":0.01198406,"domain_scores_codex":[0.9983441,0.0005924914,0.00007154192,0.0004581489,0.0004103802,0.0001234317],"domain_scores_gemma":[0.9956416,0.002042291,0.0002878441,0.00121439,0.0006284515,0.0001853938],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00009675703,0.00007987524,0.0008364597,0.0001306011,0.00007806214,0.0000989495,0.0002291031,0.7642931,0.002997769,0.1355949,0.002997479,0.09256693],"study_design_scores_gemma":[0.000004086967,0.000009832629,0.00003422593,0.000005445474,0.000004547192,0.00001564803,0.000008982327,0.9749495,0.000699477,0.02359524,0.0006662983,0.000006756017],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00553134,0.0001587851,0.9924443,0.0001781286,0.00003301086,0.00003272153,0.0001128669,0.0007050357,0.0008036736],"genre_scores_gemma":[0.388942,0.0006893813,0.5988596,0.0004419587,0.0001343441,0.000230753,0.0009906252,0.000891677,0.008819647],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007679496,"threshold_uncertainty_score":0.01526958,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2091214390","doi":"10.1109/tkde.2012.220","title":"Bias Correction in a Small Sample from Big Data","year":2012,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":71,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Estimator; Sample size determination; Computer science; Sampling bias; Big data; Sample (material); Simple random sample; Population size; Sampling (signal processing); Statistics; Reciprocal; Population; Point estimation; Data mining; Mathematics; Telecommunications","authors":[{"name":"Jianguo Lü","is_ca":true},{"name":"Dingding Li","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1054743786026927,"gpt":0.291973674795889,"spread":0.1864992961931962,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05336294,0.0009854879,0.00174783,0.002073823,0.001121138,0.002306399,0.002341002,0.001940022,0.001703827],"category_scores_gemma":[0.2569131,0.0008615612,0.001024927,0.002924754,0.002859082,0.004554228,0.002894252,0.002922851,0.0006555266],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001225923,"about_ca_system_score_gemma":0.001474776,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002482749,"about_ca_topic_score_gemma":0.001973321,"domain_scores_codex":[0.9725424,0.01825464,0.001158369,0.003457364,0.004024589,0.0005627336],"domain_scores_gemma":[0.8127549,0.1548711,0.007563912,0.01717614,0.006913932,0.0007199163],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001083659,0.0002604633,0.06111751,0.001768613,0.00161141,0.001844329,0.002093448,0.1871778,0.005235617,0.3958166,0.01871389,0.3232767],"study_design_scores_gemma":[0.0001958502,0.0002881341,0.01055522,0.0003641165,0.0002723013,0.000863247,0.0003766689,0.5855778,0.006137665,0.3788186,0.01642889,0.0001215587],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01303547,0.001106972,0.9824145,0.001067668,0.000458809,0.0001576751,0.0001773003,0.0003687177,0.001212827],"genre_scores_gemma":[0.5584185,0.001590315,0.4293938,0.003085807,0.001200574,0.001028674,0.0008122456,0.0003820747,0.004088202],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05336294,"threshold_uncertainty_score":0.2822136,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2076039115","doi":"10.1109/tkde.2009.25","title":"Evaluating the Generation of Domain Ontologies in the Knowledge Puzzle Project","year":2009,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Topic Modeling","field":"Computer Science","cited_by":69,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Ontology; Computer science; Ontology learning; Upper ontology; Information retrieval; Domain (mathematical analysis); Ontology-based data integration; Domain knowledge; Process ontology; Suggested Upper Merged Ontology; Set (abstract data type); Ontology alignment; Natural language processing; Data science; Artificial intelligence; Semantic Web","authors":[{"name":"Amal Zouaq","is_ca":true},{"name":"Roger Nkambou","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1546157879818343,"gpt":0.3661375498068649,"spread":0.2115217618250305,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01135277,0.001083707,0.0008952143,0.002574671,0.0008788874,0.002325331,0.001997393,0.001963608,0.002879117],"category_scores_gemma":[0.06540478,0.0004931436,0.00058971,0.002676472,0.00118307,0.004852819,0.003503022,0.001240607,0.0007559828],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001825307,"about_ca_system_score_gemma":0.001643117,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005153508,"about_ca_topic_score_gemma":0.005120262,"domain_scores_codex":[0.9895079,0.005993048,0.0008969897,0.001039462,0.002334165,0.0002285652],"domain_scores_gemma":[0.941774,0.04761125,0.00151219,0.003718784,0.004642796,0.0007411059],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.004314326,0.003838238,0.01834477,0.003120194,0.0004362896,0.001627328,0.006198972,0.1660923,0.01857604,0.01264454,0.02278239,0.7420247],"study_design_scores_gemma":[0.001474552,0.003632907,0.02652038,0.0004029546,0.0003186991,0.001302909,0.006953436,0.8196582,0.07659437,0.01328424,0.04966199,0.0001954269],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8988647,0.000856367,0.07929979,0.0006158837,0.0001277892,0.001342615,0.002410328,0.004623787,0.01185873],"genre_scores_gemma":[0.6746054,0.0005144653,0.304245,0.0002202565,0.00003265066,0.001032556,0.01423028,0.0006784412,0.004440981],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01135277,"threshold_uncertainty_score":0.06003994,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2131967083","doi":"10.1109/tkde.2010.262","title":"Resilient Identity Crime Detection","year":2011,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":68,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Advantage Forensics (Canada)","funders":"Australian Research Council; University of Warwick","keywords":"Computer science; Identity (music); Computer security","authors":[{"name":"Clifton Phua","is_ca":false},{"name":"Kate Smith‐Miles","is_ca":false},{"name":"Vincent Lee","is_ca":false},{"name":"Ross W. Gayler","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03861372655258806,"gpt":0.271454297643489,"spread":0.2328405710909009,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002267516,0.0007431807,0.001104013,0.003362109,0.0008835218,0.002220087,0.001879593,0.00115905,0.001369971],"category_scores_gemma":[0.01115458,0.000438796,0.0007550623,0.001409697,0.0007824133,0.002692603,0.004092277,0.001246306,0.0008278784],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001022603,"about_ca_system_score_gemma":0.001162979,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00198873,"about_ca_topic_score_gemma":0.001564285,"domain_scores_codex":[0.9968004,0.0004901395,0.0002710189,0.0007344895,0.00131346,0.0003904373],"domain_scores_gemma":[0.9913669,0.002031716,0.001703744,0.002661683,0.001869072,0.0003668573],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007500381,0.0008569303,0.07616997,0.0002398335,0.0003768206,0.0007671678,0.0007445374,0.06164958,0.05203388,0.0118966,0.007580366,0.7869341],"study_design_scores_gemma":[0.00004392264,0.0005490647,0.02223008,0.00005465614,0.0001720541,0.001493726,0.0003895353,0.8707913,0.07682346,0.01696953,0.01033189,0.0001507479],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3630505,0.0004052595,0.6169004,0.001088838,0.0001711223,0.0004915092,0.0005381503,0.01051462,0.006839645],"genre_scores_gemma":[0.8849753,0.000117531,0.1121732,0.0002298798,0.000043334,0.00008073035,0.0003129887,0.00005685861,0.002010291],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003362109,"threshold_uncertainty_score":0.01199186,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2129500661","doi":"10.1109/tkde.2008.162","title":"Mining Projected Clusters in High-Dimensional Spaces","year":2008,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":63,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Cluster analysis; Computer science; Curse of dimensionality; Linear subspace; Outlier; Clustering high-dimensional data; Data mining; CURE data clustering algorithm; Data point; Correlation clustering; Computation; Pattern recognition (psychology); Canopy clustering algorithm; Single-linkage clustering; Algorithm; Artificial intelligence; Mathematics","authors":[{"name":"Mohamed Bouguessa","is_ca":true},{"name":"Shengrui Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03923416376241975,"gpt":0.2902071390881266,"spread":0.2509729753257069,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003843067,0.001458915,0.002320229,0.005876507,0.002084518,0.003404848,0.002761962,0.002166633,0.0009169714],"category_scores_gemma":[0.01857025,0.001056979,0.001927939,0.005983123,0.001610611,0.003401351,0.00403104,0.001920173,0.00085037],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009341603,"about_ca_system_score_gemma":0.001622829,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003376128,"about_ca_topic_score_gemma":0.003143603,"domain_scores_codex":[0.9943976,0.001717083,0.0003905838,0.001161525,0.001996562,0.0003366333],"domain_scores_gemma":[0.9909,0.003748921,0.0009833293,0.001255505,0.002773721,0.0003384296],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007366231,0.0005352815,0.02626669,0.0008861371,0.000890365,0.001501383,0.002813942,0.515168,0.009665623,0.06391372,0.01102143,0.3666008],"study_design_scores_gemma":[0.00003637567,0.00007805529,0.002673762,0.00004689402,0.00004192767,0.0002743098,0.0006276823,0.919227,0.002476842,0.0724598,0.002007036,0.00005036467],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06225463,0.0003778762,0.9349715,0.000251377,0.00004252632,0.0001841339,0.0004301528,0.0006967588,0.0007910997],"genre_scores_gemma":[0.3365365,0.0005125635,0.6580942,0.0001014015,0.00009495959,0.0004147555,0.002869759,0.0001330442,0.001242849],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005876507,"threshold_uncertainty_score":0.02032435,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2543722153","doi":"10.1109/tkde.2017.2740284","title":"Activity Maximization by Effective Information Diffusion in Social Networks","year":2017,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":63,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Computer science; Submodular set function; Scalability; Maximization; Polling; Approximation algorithm; Heuristic; Upper and lower bounds; Social network (sociolinguistics); Mathematical optimization; Algorithm; Artificial intelligence; Social media; Mathematics","authors":[{"name":"Zhefeng Wang","is_ca":false},{"name":"Yang Yu","is_ca":true},{"name":"Jian Pei","is_ca":true},{"name":"Lingyang Chu","is_ca":true},{"name":"Enhong Chen","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01055453571133445,"gpt":0.2678646365113213,"spread":0.2573101007999868,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006335279,0.001711823,0.002444399,0.002154956,0.0008388403,0.002362034,0.003056147,0.002112632,0.001828526],"category_scores_gemma":[0.02648151,0.00136988,0.001408265,0.002332972,0.003190647,0.005333093,0.002497831,0.002042774,0.000403564],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002750047,"about_ca_system_score_gemma":0.001126022,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003163791,"about_ca_topic_score_gemma":0.002950142,"domain_scores_codex":[0.9968727,0.001731395,0.000121041,0.0006649072,0.000396071,0.0002138539],"domain_scores_gemma":[0.9812148,0.01588449,0.00119559,0.000877679,0.0004948883,0.000332683],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001377481,0.00007660185,0.001586153,0.00024243,0.0001026938,0.0001022477,0.0002472738,0.8475561,0.001656148,0.1163433,0.001312319,0.03063694],"study_design_scores_gemma":[0.00001305791,0.0000161395,0.00013977,0.00001093667,0.00000855644,0.00001846903,0.00001393106,0.9514596,0.0003509022,0.04765987,0.0003001423,0.000008548936],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02267579,0.0004137595,0.9746808,0.0003930181,0.00001933807,0.00007766149,0.00008599837,0.0001800024,0.001473692],"genre_scores_gemma":[0.7317803,0.001091662,0.2621874,0.0002782984,0.0001622106,0.0004902862,0.000380798,0.0001948529,0.003434247],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006335279,"threshold_uncertainty_score":0.03350455,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2766449686","doi":"10.1109/tkde.2017.2766059","title":"A Location-Query-Browse Graph for Contextual Recommendation","year":2017,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":61,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Australian Research Council","keywords":"Computer science; Graph; Bipartite graph; Information retrieval; Graph database; Web search query; Web query classification; Homogeneous; Theoretical computer science; World Wide Web; Search engine; Mathematics; Combinatorics","authors":[{"name":"Yongli Ren","is_ca":false},{"name":"Martin Tomko","is_ca":false},{"name":"Flora D. Salim","is_ca":false},{"name":"Jeffrey Chan","is_ca":false},{"name":"Charles L. A. Clarke","is_ca":true},{"name":"Mark Sanderson","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0619002510906548,"gpt":0.3532966062404341,"spread":0.2913963551497793,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004316978,0.0007776735,0.0006905437,0.00277406,0.0009658253,0.0008583753,0.001326976,0.001164368,0.003365674],"category_scores_gemma":[0.004935626,0.0004933047,0.001033407,0.005523005,0.0005845435,0.001908369,0.000934881,0.000923437,0.001005006],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001117476,"about_ca_system_score_gemma":0.001231711,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.04347202,"about_ca_topic_score_gemma":0.08334941,"domain_scores_codex":[0.999136,0.000217898,0.0000458919,0.0003452031,0.0002064536,0.00004863704],"domain_scores_gemma":[0.9973231,0.001078925,0.0002749845,0.0007768985,0.0004033832,0.0001426826],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004624214,0.0006111236,0.05998185,0.002167301,0.0008290329,0.00155344,0.00182119,0.3448944,0.01736744,0.1229273,0.0932772,0.3541072],"study_design_scores_gemma":[0.00005973523,0.0001490509,0.01854208,0.0001964569,0.0003389926,0.001450361,0.0005846669,0.8373446,0.003385887,0.08034293,0.05749501,0.000110163],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08498084,0.002303344,0.8639768,0.001313059,0.0001314434,0.0005863325,0.03558103,0.003046922,0.008080285],"genre_scores_gemma":[0.4939828,0.001389818,0.4603088,0.0005222034,0.00008976318,0.0004941638,0.03920137,0.0002710887,0.003740049],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.04347202,"threshold_uncertainty_score":0.086438,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4400033057","doi":"10.1109/tkde.2024.3419215","title":"Ze-HFS: Zentropy-Based Uncertainty Measure for Heterogeneous Feature Selection and Knowledge Discovery","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":56,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"National Natural Science Foundation of China","keywords":"Computer science; Measure (data warehouse); Feature selection; Selection (genetic algorithm); Feature (linguistics); Knowledge extraction; Data mining; Artificial intelligence","authors":[{"name":"Kehua Yuan","is_ca":false},{"name":"Duoqian Miao","is_ca":false},{"name":"Witold Pedrycz","is_ca":true},{"name":"Weiping Ding","is_ca":false},{"name":"Hongyun Zhang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01879242851704101,"gpt":0.2692210126805711,"spread":0.2504285841635301,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00358836,0.000963534,0.001507117,0.002919973,0.0008104047,0.001656486,0.001449829,0.000876471,0.0009418677],"category_scores_gemma":[0.00921812,0.0002959797,0.001741942,0.00221957,0.001315421,0.002525741,0.001849994,0.00112976,0.000141298],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001436056,"about_ca_system_score_gemma":0.001344561,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00319151,"about_ca_topic_score_gemma":0.001867305,"domain_scores_codex":[0.9959856,0.001059379,0.0003512944,0.0006738567,0.00172192,0.0002078637],"domain_scores_gemma":[0.9965953,0.002019651,0.0003681378,0.0003360453,0.0005740916,0.0001067718],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003585668,0.0001292728,0.007662356,0.0004939198,0.000626958,0.0003515752,0.0003453685,0.4865382,0.01041224,0.09210981,0.003673603,0.3972981],"study_design_scores_gemma":[0.00002265606,0.0001098151,0.002209519,0.00003361588,0.0000690113,0.0001326511,0.00005516177,0.9450763,0.003997188,0.04639456,0.00184581,0.0000537336],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01126829,0.0004439036,0.9873075,0.0001312712,0.00002647977,0.00004321073,0.00009944229,0.0001475336,0.0005323414],"genre_scores_gemma":[0.6771804,0.0009714806,0.31855,0.0002650808,0.0002057127,0.0003628865,0.0009066692,0.000096972,0.001460923],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00358836,"threshold_uncertainty_score":0.01897728,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2894983428","doi":"10.1109/tkde.2018.2874004","title":"Precise and Fast Cryptanalysis for Bloom Filter Based Privacy-Preserving Record Linkage","year":2018,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":56,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Australian Research Council; Engineering and Physical Sciences Research Council; Simons Foundation","keywords":"Bloom filter; Computer science; Cryptanalysis; Encoding (memory); Scalability; Data mining; Filter (signal processing); Information retrieval; Algorithm; Database; Cryptography; Artificial intelligence","authors":[{"name":"Peter Christen","is_ca":false},{"name":"Thilina Ranbaduge","is_ca":false},{"name":"Dinusha Vatsalan","is_ca":false},{"name":"Rainer Schnell","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1370625278578638,"gpt":0.3797145295626111,"spread":0.2426520017047473,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005850377,0.001108772,0.001572477,0.002319164,0.001980615,0.003609297,0.002059967,0.002419779,0.002685955],"category_scores_gemma":[0.02087376,0.0007535715,0.001436072,0.003275026,0.00231772,0.009522171,0.004214883,0.00317126,0.001867294],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003027432,"about_ca_system_score_gemma":0.003981203,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001207124,"about_ca_topic_score_gemma":0.0008897713,"domain_scores_codex":[0.9827316,0.004171852,0.00149133,0.001565099,0.008570208,0.001469955],"domain_scores_gemma":[0.9755321,0.009919424,0.002910034,0.008930675,0.002313276,0.0003945783],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.003054492,0.0007620446,0.008067505,0.0007431131,0.0005532136,0.001043398,0.00150849,0.09085476,0.1211621,0.1693958,0.0162141,0.586641],"study_design_scores_gemma":[0.0002843579,0.0006212834,0.001392166,0.0002158122,0.0001726644,0.002172864,0.0003832046,0.6150941,0.207863,0.14529,0.02627412,0.0002363723],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05367023,0.001156429,0.9329184,0.001307732,0.0001693658,0.0004064108,0.0004310946,0.005563008,0.004377346],"genre_scores_gemma":[0.6087893,0.0006618716,0.3839929,0.0007536131,0.0001160585,0.0003894629,0.0005912464,0.0002663401,0.00443922],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005850377,"threshold_uncertainty_score":0.03094012,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2128933128","doi":"10.1109/tkde.2008.38","title":"Simultaneous Pattern and Data Clustering for Pattern Cluster Analysis","year":2008,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":55,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Data mining; Cluster analysis; Categorical variable; Relation (database); Cluster (spacecraft); Data set; Set (abstract data type); Pattern recognition (psychology); Consensus clustering; Measure (data warehouse); Artificial intelligence; Fuzzy clustering; CURE data clustering algorithm; Machine learning","authors":[{"name":"Andrew K. C. Wong","is_ca":true},{"name":"G.C.L. Li","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03944602358900941,"gpt":0.2853976676282162,"spread":0.2459516440392067,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005375003,0.00257738,0.002355215,0.005952595,0.002237939,0.002918095,0.003249121,0.002132346,0.005925441],"category_scores_gemma":[0.01842988,0.0009902064,0.002908827,0.01249242,0.002107578,0.004704318,0.004339161,0.00391329,0.00301008],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001336112,"about_ca_system_score_gemma":0.002590618,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002942312,"about_ca_topic_score_gemma":0.003081787,"domain_scores_codex":[0.9891651,0.003977937,0.0007816641,0.002444285,0.0032921,0.0003388532],"domain_scores_gemma":[0.9919435,0.003356928,0.0006361281,0.002329498,0.001575132,0.0001588444],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003539117,0.0002392073,0.002766054,0.001564502,0.0007248315,0.0004305415,0.00108513,0.06390357,0.009194709,0.1477842,0.01842519,0.7535281],"study_design_scores_gemma":[0.00006922565,0.0001517608,0.001543775,0.0001648066,0.0001855711,0.0007549088,0.0003024196,0.6809596,0.01287672,0.234316,0.06850021,0.0001749764],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0009133389,0.0003880927,0.9967086,0.0001382165,0.00007008675,0.0001215392,0.0001258526,0.0009001699,0.0006340228],"genre_scores_gemma":[0.02365411,0.0004507854,0.9734762,0.0001405018,0.0000987626,0.0005240075,0.0006247838,0.0002254868,0.0008053708],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005952595,"threshold_uncertainty_score":0.02842605,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3167049326","doi":"10.1109/tkde.2023.3328882","title":"DASVDD: Deep Autoencoding Support Vector Data Descriptor for Anomaly Detection","year":2023,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":54,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Mila - Quebec Artificial Intelligence Institute","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Anomaly detection; Support vector machine; Artificial intelligence; Pattern recognition (psychology); Anomaly (physics); Data mining; Data modeling","authors":[{"name":"Hadi Hojjati","is_ca":true},{"name":"Narges Armanfard","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05801809301462153,"gpt":0.298694577054953,"spread":0.2406764840403315,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008499808,0.0008297375,0.001137397,0.001370124,0.0002605196,0.000999077,0.001558523,0.0007147422,0.001310738],"category_scores_gemma":[0.002765095,0.0003004949,0.0008239795,0.001251503,0.000541766,0.001394714,0.001381215,0.001669937,0.0006977547],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007342415,"about_ca_system_score_gemma":0.001132481,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003871246,"about_ca_topic_score_gemma":0.003907304,"domain_scores_codex":[0.9992619,0.0001058447,0.000056429,0.0001686043,0.0003279882,0.00007926204],"domain_scores_gemma":[0.9990994,0.0002514796,0.0001176656,0.0001891796,0.0002815122,0.00006071829],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000262463,0.0001732153,0.005045607,0.0001504516,0.0001362124,0.000134954,0.00006691235,0.1783829,0.01381247,0.009024957,0.0139137,0.7788962],"study_design_scores_gemma":[0.000007795094,0.0000313885,0.0004236123,0.000006384427,0.000005894155,0.00005654841,0.00001384055,0.9892269,0.005023779,0.003598226,0.001593614,0.00001195051],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01728626,0.0004280815,0.9790071,0.0001580862,0.0001092177,0.00004515355,0.0004576951,0.00199095,0.0005173338],"genre_scores_gemma":[0.5257683,0.0006880765,0.4630404,0.0003018586,0.000102769,0.0001931271,0.004976867,0.0002798477,0.004648691],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003871246,"threshold_uncertainty_score":0.007697403,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3163325638","doi":"10.1109/tkde.2021.3075052","title":"Learning Hierarchical Review Graph Representations for Recommendation","year":2021,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":51,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"BC Research (Canada)","funders":"Youth Innovation Promotion Association; Youth Innovation Promotion Association of the Chinese Academy of Sciences; National Research Foundation Singapore; National Natural Science Foundation of China; Nanyang Technological University; National Research Foundation","keywords":"Computer science; Pooling; Graph; Recommender system; Artificial intelligence; Convolutional neural network; Recurrent neural network; Machine learning; Information retrieval; Theoretical computer science; Artificial neural network","authors":[{"name":"Yong Liu","is_ca":true},{"name":"Susen Yang","is_ca":true},{"name":"Yinan Zhang","is_ca":true},{"name":"Chunyan Miao","is_ca":true},{"name":"Zaiqing Nie","is_ca":true},{"name":"Juyong Zhang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04649403019805645,"gpt":0.3242526013051148,"spread":0.2777585711070584,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000664695,0.0008747086,0.0008660458,0.001806339,0.000298427,0.0007107144,0.001085681,0.0009535318,0.001994242],"category_scores_gemma":[0.004877248,0.0004130192,0.0009169254,0.002111702,0.0002813181,0.002039811,0.0005098816,0.001190876,0.001193557],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008441636,"about_ca_system_score_gemma":0.0006838197,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01365356,"about_ca_topic_score_gemma":0.03306761,"domain_scores_codex":[0.9994608,0.0001552187,0.00003279967,0.000202697,0.0001126692,0.00003590137],"domain_scores_gemma":[0.9985708,0.0005998067,0.0002172312,0.0002646816,0.0003002475,0.00004717723],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003686855,0.0002770057,0.007181252,0.0006361788,0.0004768585,0.0001928294,0.0002481356,0.2521788,0.01122319,0.01923596,0.02237755,0.6856034],"study_design_scores_gemma":[0.00001196157,0.0000656113,0.001381441,0.0000212465,0.00006854426,0.00006322197,0.00001955116,0.9811101,0.001214986,0.01265186,0.00337237,0.00001916445],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06051541,0.004587523,0.9245968,0.0006269384,0.0001579169,0.0001905548,0.002251656,0.003022492,0.004050657],"genre_scores_gemma":[0.6900629,0.002904952,0.287755,0.0004989982,0.0002515714,0.0002916246,0.007545302,0.0002384624,0.01045114],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01365356,"threshold_uncertainty_score":0.02714813,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2127426313","doi":"10.1109/69.929898","title":"Constructing the dependency structure of a multiagent probabilistic network","year":2001,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Bayesian Modeling and Causal Inference","field":"Computer Science","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Regina","funders":"","keywords":"Dependency (UML); Probabilistic logic; Computer science; Cover (algebra); Graphical model; Representation (politics); Theoretical computer science; Domain (mathematical analysis); Dependency graph; Conditional probability; Artificial intelligence; Data mining; Mathematics; Graph","authors":[{"name":"S. K. M. Wong","is_ca":true},{"name":"Cory J. Butz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0310304745145064,"gpt":0.2611466064437121,"spread":0.2301161319292057,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001229781,0.0004910256,0.0004468837,0.001852038,0.0008444364,0.0008678624,0.0008681125,0.0007945161,0.002868031],"category_scores_gemma":[0.00681308,0.0007002478,0.001125389,0.0009912523,0.001018552,0.002831173,0.001731937,0.001043992,0.0004476767],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00114968,"about_ca_system_score_gemma":0.001116163,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003634576,"about_ca_topic_score_gemma":0.003802836,"domain_scores_codex":[0.9985903,0.0004361602,0.00008002682,0.0003978226,0.0004164667,0.00007922116],"domain_scores_gemma":[0.9967669,0.002010412,0.0002827664,0.0004559265,0.0003957282,0.00008829095],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002119202,0.00008237459,0.008263402,0.0003219383,0.0002034209,0.001277544,0.00094394,0.4794039,0.009241381,0.368111,0.005070075,0.126869],"study_design_scores_gemma":[0.00001388126,0.00002506105,0.00160554,0.00003313114,0.00005573219,0.0001780818,0.00005266473,0.7768009,0.003555486,0.2126045,0.005050987,0.00002397785],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02108186,0.00005691047,0.9759136,0.0002108158,0.000007030055,0.00005011915,0.0004174551,0.0002751705,0.001987124],"genre_scores_gemma":[0.4746268,0.0002591506,0.5205736,0.0001019348,0.00003226506,0.0002992645,0.002215296,0.0001050219,0.001786651],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003634576,"threshold_uncertainty_score":0.0095945,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2807633791","doi":"10.1109/tkde.2018.2857471","title":"Secure and Efficient Skyline Queries on Encrypted Data","year":2018,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Cryptography and Data Security","field":"Computer Science","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"National Institutes of Health; National Institute of General Medical Sciences; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Patient-Centered Outcomes Research Institute","keywords":"Computer science; Encryption; Cloud computing; Skyline; Scalability; Database; Outsourcing; Computer network; Data mining; Operating system","authors":[{"name":"Jinfei Liu","is_ca":false},{"name":"Juncheng Yang","is_ca":false},{"name":"Li Xiong","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02885223219186742,"gpt":0.2747359068435998,"spread":0.2458836746517324,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003028712,0.0006326549,0.001178503,0.0006404503,0.00107117,0.002398516,0.001303064,0.0009972113,0.001894841],"category_scores_gemma":[0.007298469,0.0003551158,0.0008133309,0.001257515,0.001254633,0.005943774,0.003284022,0.001192425,0.0007991684],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00113416,"about_ca_system_score_gemma":0.001647735,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001270414,"about_ca_topic_score_gemma":0.001337854,"domain_scores_codex":[0.9950929,0.001314816,0.0005118286,0.0005985305,0.001748319,0.0007336594],"domain_scores_gemma":[0.9931436,0.002386926,0.0009051582,0.002453963,0.0009103302,0.0002001542],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.003676281,0.0005608113,0.005637582,0.0005793499,0.0003102488,0.001272202,0.002856566,0.3116879,0.1823898,0.2595067,0.01361629,0.2179064],"study_design_scores_gemma":[0.0002170241,0.0004022804,0.0009724919,0.00003073443,0.00004472583,0.0005183064,0.0008055008,0.8228993,0.06958849,0.09675914,0.007707113,0.00005481993],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1514979,0.0002967832,0.8422624,0.0006911202,0.00003200763,0.0002960318,0.0006313878,0.001097796,0.003194551],"genre_scores_gemma":[0.8380193,0.000250483,0.1580249,0.0001442209,0.00005103223,0.0002377427,0.0007288422,0.0001404567,0.002402951],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003028712,"threshold_uncertainty_score":0.0160175,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2110420974","doi":"10.1109/tkde.2010.186","title":"Exact Top-K Queries in Wireless Sensor Networks","year":2010,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Energy Efficient Wireless Sensor Networks","field":"Computer Science","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Network topology; Wireless sensor network; Context (archaeology); Tree (set theory); Topology (electrical circuits); Algorithm; Set (abstract data type); Logical topology; Theoretical computer science; Computer network; Mathematics","authors":[{"name":"Baljeet Malhotra","is_ca":true},{"name":"Mário A. Nascimento","is_ca":true},{"name":"Ioanis Nikolaidis","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01204658840230878,"gpt":0.2362237369969768,"spread":0.224177148594668,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005270128,0.001361019,0.003380544,0.001579559,0.002214561,0.003488054,0.003497911,0.00248267,0.001678265],"category_scores_gemma":[0.01909655,0.0008448166,0.0009491954,0.004066952,0.002307876,0.01037602,0.003016823,0.00165109,0.000747418],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001228591,"about_ca_system_score_gemma":0.0018206,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002858444,"about_ca_topic_score_gemma":0.003123222,"domain_scores_codex":[0.9946792,0.001362905,0.0005877683,0.001260717,0.001609238,0.0005002159],"domain_scores_gemma":[0.9838324,0.01048777,0.001301768,0.003023592,0.001012165,0.0003423115],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002079334,0.000280587,0.008136188,0.001130822,0.0003501472,0.0006708052,0.000925385,0.6422607,0.01408718,0.04196185,0.01218963,0.2759274],"study_design_scores_gemma":[0.00006566426,0.0001661619,0.0006717815,0.0000287253,0.00005071146,0.0005683939,0.0003710093,0.9103984,0.00579366,0.07954909,0.00228725,0.000049166],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08190465,0.003137994,0.9088887,0.001057234,0.0001728423,0.000189942,0.000757344,0.002093894,0.00179742],"genre_scores_gemma":[0.6944894,0.001096154,0.3010228,0.0004803364,0.0001910951,0.0001636174,0.0009979985,0.0002129241,0.001345729],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005270128,"threshold_uncertainty_score":0.02787143,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2003787915","doi":"10.1109/tkde.2014.2310219","title":"Discovery of Temporal Associations in Multivariate Time Series","year":2014,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Time Series Analysis and Forecasting","field":"Computer Science","cited_by":47,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Data mining; Multivariate statistics; Time series; Pruning; Scalability; Redundancy (engineering); Series (stratigraphy); Artificial intelligence; Machine learning","authors":[{"name":"Dennis Zhuang","is_ca":true},{"name":"Gary C.L. Li","is_ca":true},{"name":"Andrew K. C. Wong","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01599466251246835,"gpt":0.2367904487629129,"spread":0.2207957862504445,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001522327,0.0006840869,0.0008588765,0.003489839,0.0005214961,0.0009756734,0.00068836,0.0004605504,0.0005719177],"category_scores_gemma":[0.008611109,0.0003128063,0.0008574013,0.004889132,0.0003796103,0.001404953,0.0009800222,0.0009087472,0.0002340198],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002912756,"about_ca_system_score_gemma":0.0006407764,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002084514,"about_ca_topic_score_gemma":0.003166917,"domain_scores_codex":[0.9987362,0.0001940975,0.0001358031,0.0003304659,0.0004902551,0.0001131444],"domain_scores_gemma":[0.9954002,0.002312486,0.00115784,0.0004523808,0.0005371135,0.0001399065],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005819089,0.0004342168,0.1487457,0.0004370781,0.0004451615,0.002083504,0.0008920666,0.1718499,0.02925279,0.02352619,0.003600565,0.6181509],"study_design_scores_gemma":[0.0000139481,0.00009509289,0.02467389,0.00002869554,0.00007848774,0.0005007396,0.000173095,0.9502528,0.004261914,0.01667311,0.00321637,0.00003190058],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2533352,0.0005983574,0.7434038,0.000194111,0.00006083487,0.00007463346,0.0007015392,0.0005946676,0.001036861],"genre_scores_gemma":[0.8002976,0.0005281142,0.1964441,0.00005196581,0.0001372469,0.0001073781,0.001656744,0.00006108185,0.0007157385],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003489839,"threshold_uncertainty_score":0.008050978,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3128735821","doi":"10.1109/tkde.2021.3054671","title":"Transfer Learning for Dynamic Feature Extraction Using Variational Bayesian Inference","year":2021,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Fault Detection and Control Systems","field":"Engineering","cited_by":47,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Inference; Bayesian inference; Weighting; Feature (linguistics); Machine learning; Transfer of learning; Artificial intelligence; Data mining; Bayesian probability; Domain (mathematical analysis); Mathematics","authors":[{"name":"Junyao Xie","is_ca":true},{"name":"Biao Huang","is_ca":true},{"name":"Stevan Dubljević","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01656466914588841,"gpt":0.2718116489394927,"spread":0.2552469797936043,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002615644,0.001166678,0.001363697,0.001288548,0.0005192772,0.001114937,0.002161351,0.001471478,0.002701658],"category_scores_gemma":[0.007393206,0.001013521,0.001663777,0.00126406,0.001188444,0.001885062,0.002178943,0.002693261,0.0005656411],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001363885,"about_ca_system_score_gemma":0.001551422,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008378936,"about_ca_topic_score_gemma":0.006406134,"domain_scores_codex":[0.9992092,0.0002734325,0.00004800809,0.0001904578,0.0001977527,0.00008120042],"domain_scores_gemma":[0.9975757,0.001902999,0.0001332929,0.0001094065,0.0002324761,0.00004614254],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000449515,0.00004735298,0.0004642849,0.00007921138,0.00006469584,0.00005700704,0.00007015337,0.8954777,0.001493149,0.02091551,0.0008803271,0.08040562],"study_design_scores_gemma":[0.000001973974,0.000004290163,0.00002782584,0.000002219649,0.000002066897,0.000003697505,0.000001599107,0.9947082,0.0001448475,0.00498261,0.0001178792,0.000002773707],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001512161,0.00007069647,0.9980029,0.00004771939,0.000006530469,0.00001627813,0.00002310514,0.000114813,0.0002057435],"genre_scores_gemma":[0.5020055,0.0006293199,0.4913394,0.0002465135,0.0001196159,0.0006138463,0.0006957216,0.0003089919,0.004041085],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008378936,"threshold_uncertainty_score":0.01666033,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2134465446","doi":"10.1109/tkde.2006.11","title":"Input variable selection: mutual information and linear mixing measures","year":2006,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Blind Source Separation Techniques","field":"Computer Science","cited_by":41,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Independent component analysis; Mutual information; Computer science; Preprocessor; Dependency (UML); Data pre-processing; Data mining; Mixing (physics); Feature selection; Pattern recognition (psychology); Blind signal separation; Algorithm; Gaussian; Artificial intelligence; Channel (broadcasting)","authors":[{"name":"Thomas Trappenberg","is_ca":true},{"name":"Jie Ouyang","is_ca":false},{"name":"Andrew D. Back","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01617825178139467,"gpt":0.2422223643631334,"spread":0.2260441125817387,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00645346,0.001111001,0.001024343,0.002083347,0.0006129742,0.002191385,0.001234547,0.001704916,0.001229323],"category_scores_gemma":[0.02814614,0.0004072277,0.0008457256,0.001418346,0.001998764,0.003080239,0.001873758,0.001402825,0.0002843376],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009140504,"about_ca_system_score_gemma":0.0009382599,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007006774,"about_ca_topic_score_gemma":0.0005052219,"domain_scores_codex":[0.9959371,0.00193204,0.000251323,0.0006399095,0.001067388,0.0001722393],"domain_scores_gemma":[0.9836767,0.0130647,0.001323218,0.0008547797,0.0008816354,0.0001989891],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007859314,0.0003114353,0.01435234,0.0004878475,0.0005093671,0.0002686562,0.000306101,0.5152358,0.02245652,0.1340138,0.002423245,0.3088489],"study_design_scores_gemma":[0.00002270102,0.0001296793,0.00233501,0.00003126285,0.00005123165,0.00009347305,0.00002979941,0.9376958,0.01095846,0.04787211,0.0007351019,0.00004537687],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02357611,0.0003218394,0.9740716,0.0002452199,0.00001774718,0.00005448699,0.00007869546,0.0003110074,0.001323351],"genre_scores_gemma":[0.6299979,0.0003509371,0.3674946,0.0001663455,0.0001418757,0.0003014824,0.0003759799,0.0002132243,0.0009577162],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00645346,"threshold_uncertainty_score":0.03412956,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4312529005","doi":"10.1109/tkde.2022.3221316","title":"DMGAN: Dynamic Multi-Hop Graph Attention Network for Traffic Forecasting","year":2022,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Traffic Prediction and Management Techniques","field":"Engineering","cited_by":40,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"National Natural Science Foundation of China","keywords":"Computer science; Leverage (statistics); Data mining; Graph; Network topology; Intelligent transportation system; Theoretical computer science; Artificial intelligence; Computer network","authors":[{"name":"Rui Li","is_ca":false},{"name":"Fan Zhang","is_ca":false},{"name":"Tong Li","is_ca":false},{"name":"Tingting Zhang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02779923458341519,"gpt":0.2480279719790504,"spread":0.2202287373956352,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005095306,0.001301878,0.000679932,0.001252816,0.0003544203,0.0004301862,0.00146927,0.0008675341,0.001604263],"category_scores_gemma":[0.001545024,0.0004083303,0.0006934647,0.00102262,0.0003522188,0.001311159,0.0008686446,0.001221621,0.0005116928],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001425997,"about_ca_system_score_gemma":0.0007570209,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02249759,"about_ca_topic_score_gemma":0.03177883,"domain_scores_codex":[0.9997737,0.00005428307,0.000007734317,0.00008419563,0.00003828861,0.00004179808],"domain_scores_gemma":[0.9996069,0.0002113709,0.00003742817,0.00005144209,0.00007120014,0.00002157519],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001984905,0.0001467464,0.002627156,0.0001402594,0.0001699439,0.0001565683,0.0001045033,0.5959249,0.005948281,0.008246978,0.01691041,0.3694259],"study_design_scores_gemma":[0.000004094887,0.00001454374,0.0003257945,0.000005417615,0.00001319697,0.00002107806,0.000008765658,0.9937196,0.0009337654,0.00397662,0.0009714305,0.000005660344],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06527487,0.002495989,0.9162934,0.00117128,0.0003065113,0.0001244427,0.001760979,0.007293635,0.005278778],"genre_scores_gemma":[0.7899227,0.001377737,0.1907717,0.001061536,0.0002671331,0.0001676956,0.005844566,0.0003597395,0.01022717],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02249759,"threshold_uncertainty_score":0.04473329,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4312251623","doi":"10.1109/tkde.2022.3231929","title":"Multi-View Fuzzy Classification With Subspace Clustering and Information Granules","year":2022,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Fuzzy Logic and Control Systems","field":"Computer Science","cited_by":39,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Natural Science Foundation for Distinguished Young Scholars of Hunan Province; National Natural Science Foundation of China","keywords":"Computer science; Interpretability; Cluster analysis; Subspace topology; Data mining; Fuzzy logic; Artificial intelligence; Machine learning; Fuzzy clustering","authors":[{"name":"Xingchen Hu","is_ca":false},{"name":"Xinwang Liu","is_ca":false},{"name":"Witold Pedrycz","is_ca":true},{"name":"Qing Liao","is_ca":false},{"name":"Yinghua Shen","is_ca":false},{"name":"Yan Li","is_ca":false},{"name":"Siwei Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02701039006990769,"gpt":0.234104556318824,"spread":0.2070941662489163,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001204435,0.0007495061,0.001392035,0.001700216,0.0005997485,0.001488768,0.001295048,0.001191983,0.0007846424],"category_scores_gemma":[0.00276511,0.000338411,0.001539575,0.001779547,0.0007485606,0.002001084,0.001145367,0.001066655,0.0002824113],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007508375,"about_ca_system_score_gemma":0.0007208123,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004306599,"about_ca_topic_score_gemma":0.002771792,"domain_scores_codex":[0.9990085,0.0002385051,0.0000861567,0.0002777425,0.0003059097,0.00008323784],"domain_scores_gemma":[0.9991667,0.0002855779,0.0001205058,0.0001483592,0.0002367926,0.00004199311],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002514875,0.0001236576,0.00283687,0.0002197445,0.0002462719,0.0002069058,0.0003896862,0.5547646,0.008169776,0.04317282,0.002752941,0.3868652],"study_design_scores_gemma":[0.000003854268,0.00001755824,0.0002198799,0.00000639359,0.0000108742,0.00001892063,0.00002046735,0.9899893,0.0007776363,0.008632204,0.000293973,0.000008786736],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01055855,0.0002021841,0.9885923,0.00006475436,0.00001803484,0.00002509776,0.00003998347,0.0001294433,0.0003697689],"genre_scores_gemma":[0.5569754,0.000407903,0.4407297,0.0000963186,0.0000931057,0.0001576963,0.0004692956,0.00004398849,0.001026613],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004306599,"threshold_uncertainty_score":0.008563101,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2796213239","doi":"10.1109/tkde.2018.2821671","title":"Characterizing and Predicting Early Reviewers for Effective Product Marketing on E-Commerce Websites","year":2018,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Digital Marketing and Social Media","field":"Social Sciences","cited_by":37,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"National Key Research and Development Program of China; Innovate UK; Natural Science Foundation of Beijing Municipality; Renmin University of China; National Natural Science Foundation of China","keywords":"Helpfulness; Popularity; Product (mathematics); Computer science; New product development; Information retrieval; Artificial intelligence; World Wide Web; Psychology; Marketing; Mathematics; Business; Social psychology","authors":[{"name":"Wayne Xin Zhao","is_ca":false},{"name":"Yulan He","is_ca":false},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Ji-Rong Wen","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02389185359157603,"gpt":0.3001767249798764,"spread":0.2762848713883004,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007456454,0.0007409006,0.0006353726,0.004197037,0.00101163,0.002865233,0.0007964624,0.00133673,0.001844861],"category_scores_gemma":[0.07583565,0.0004895712,0.0004622531,0.001794666,0.0004923421,0.002787923,0.00075299,0.0009170587,0.001375757],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007658376,"about_ca_system_score_gemma":0.001152985,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003253844,"about_ca_topic_score_gemma":0.008280749,"domain_scores_codex":[0.9939889,0.002106152,0.0006701814,0.00107112,0.001762211,0.0004014052],"domain_scores_gemma":[0.8452938,0.07274114,0.04543535,0.003284122,0.02576157,0.007483969],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003730094,0.0002694787,0.9565096,0.0002242775,0.0001082448,0.000330662,0.0008202086,0.001503763,0.00223484,0.0003762121,0.002719223,0.03453058],"study_design_scores_gemma":[0.00004114044,0.000642703,0.9422721,0.00009576145,0.0002010163,0.001069092,0.001321843,0.04424616,0.004290935,0.0008799171,0.004833684,0.0001056244],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9902463,0.001838399,0.004548497,0.0003977214,0.00007115436,0.0001540908,0.0004177641,0.0001084503,0.002217649],"genre_scores_gemma":[0.9932697,0.0004714009,0.003817499,0.00005254969,0.0001015396,0.00005683462,0.0004062215,0.00002776727,0.001796525],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.007456454,"threshold_uncertainty_score":0.03943396,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2160172686","doi":"10.1109/tkde.2009.138","title":"Credibility: How Agents Can Handle Unfair Third-Party Testimonies in Computational Trust Models","year":2009,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Access Control and Trust","field":"Social Sciences","cited_by":34,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Credibility; Computer science; Computer security; Work (physics); Computational trust; Empirical evidence; Computational model; Risk analysis (engineering); Artificial intelligence; Reputation; Business; Political science; Law","authors":[{"name":"Jianshu Weng","is_ca":false},{"name":"Zhiqi Shen","is_ca":false},{"name":"Chunyan Miao","is_ca":false},{"name":"Angela Goh","is_ca":false},{"name":"Cyril Leung","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05783747098160918,"gpt":0.3104771853336928,"spread":0.2526397143520836,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01222793,0.001108398,0.001833408,0.001710739,0.001950543,0.005861089,0.003571196,0.004880568,0.004600425],"category_scores_gemma":[0.0731471,0.001012867,0.001607025,0.00159792,0.003474248,0.01107495,0.005701978,0.004563733,0.0008605886],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002125808,"about_ca_system_score_gemma":0.00197235,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004556961,"about_ca_topic_score_gemma":0.002711923,"domain_scores_codex":[0.990311,0.005930192,0.0005941775,0.0009637335,0.001674196,0.0005266239],"domain_scores_gemma":[0.9534686,0.0346767,0.00314623,0.004419227,0.003269621,0.001019737],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001673744,0.00006473147,0.001254962,0.000178208,0.0001272498,0.0005607968,0.0011494,0.3392533,0.0006730929,0.6230755,0.00280576,0.03068968],"study_design_scores_gemma":[0.00003695814,0.00003314451,0.0001036149,0.00003674604,0.00003677366,0.0001323054,0.000080428,0.7189458,0.0003501645,0.277625,0.002591696,0.00002739919],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01274554,0.0003021939,0.976987,0.002140212,0.00008319476,0.000102515,0.00007966561,0.000208732,0.007350964],"genre_scores_gemma":[0.7321249,0.0007398035,0.258492,0.0004847125,0.0002686614,0.0003656109,0.0002144664,0.0001217549,0.007188144],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01222793,"threshold_uncertainty_score":0.0646683,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2784105835","doi":"10.1109/tkde.2018.2793862","title":"A Two-Phase Algorithm for Differentially Private Frequent Subgraph Mining","year":2018,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"International Development Research Centre","funders":"National Institutes of Health; National Institute of General Medical Sciences; National Natural Science Foundation of China; Deutsche Forschungsgemeinschaft; Patient-Centered Outcomes Research Institute","keywords":"Computer science; Differential privacy; Support vector machine; Pruning; Data mining; Graph; Algorithm; Theoretical computer science; Artificial intelligence","authors":[{"name":"Xiang Cheng","is_ca":false},{"name":"Sen Su","is_ca":false},{"name":"Shengzhi Xu","is_ca":true},{"name":"Li Xiong","is_ca":false},{"name":"Ke Xiao","is_ca":false},{"name":"Mingxing Zhao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03871756589284604,"gpt":0.3077567980364951,"spread":0.2690392321436491,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003269779,0.0008524488,0.001247405,0.001851451,0.001073364,0.001320412,0.002848475,0.001699951,0.002063695],"category_scores_gemma":[0.01620085,0.0005010224,0.001332848,0.002488927,0.001077179,0.003017615,0.003480574,0.001796874,0.0008521017],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001181078,"about_ca_system_score_gemma":0.002647064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001127965,"about_ca_topic_score_gemma":0.001810236,"domain_scores_codex":[0.9957729,0.00124019,0.0003662546,0.0008457169,0.001479618,0.0002953327],"domain_scores_gemma":[0.9925324,0.003733408,0.0005848359,0.002194907,0.0007310361,0.0002233555],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001445448,0.0006717845,0.007160991,0.0003412129,0.0002203856,0.00056239,0.0005608528,0.171405,0.02063222,0.08366254,0.009716583,0.7036205],"study_design_scores_gemma":[0.0001853651,0.0001572963,0.0004748497,0.00001710918,0.00003319716,0.0004988054,0.00007734701,0.8968024,0.006621017,0.09195329,0.003154266,0.00002498137],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01308808,0.0001001363,0.9848108,0.0003250996,0.00002741295,0.0002100559,0.0002462114,0.0007317441,0.0004605068],"genre_scores_gemma":[0.2472074,0.0001207621,0.7490875,0.0002619501,0.00006693995,0.0006055863,0.001336623,0.0001147459,0.001198613],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003269779,"threshold_uncertainty_score":0.01729244,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2142595897","doi":"10.1109/tkde.2006.146","title":"On the Signature Tree Construction and Analysis","year":2006,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Winnipeg","funders":"","keywords":"Computer science; Signature (topology); Tree (set theory); Set (abstract data type); Digital signature; Data structure; Data mining; File format; Information retrieval; Theoretical computer science; Database; Programming language; Hash function; Mathematics","authors":[{"name":"Yangjun Chen","is_ca":true},{"name":"Yibin Chen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01171584614459828,"gpt":0.2185866327377866,"spread":0.2068707865931883,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002052126,0.0008683396,0.001156844,0.004942249,0.001540506,0.002933491,0.001988245,0.001178056,0.004144575],"category_scores_gemma":[0.01257204,0.0006516917,0.001229575,0.009070652,0.001689327,0.008252941,0.002303738,0.002337043,0.002916873],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001233792,"about_ca_system_score_gemma":0.002445151,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002886492,"about_ca_topic_score_gemma":0.001861493,"domain_scores_codex":[0.9968713,0.000726717,0.0002824432,0.0004957417,0.001423278,0.00020045],"domain_scores_gemma":[0.9925044,0.002817229,0.0006533103,0.001844541,0.001906819,0.0002737359],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001930516,0.0001460903,0.002269823,0.0003690015,0.00006202583,0.0003378647,0.0004442112,0.03582958,0.01194742,0.2646691,0.01465295,0.6690788],"study_design_scores_gemma":[0.00005293134,0.000200827,0.001118241,0.0001322332,0.00007131648,0.001342654,0.0001997044,0.4917144,0.01675718,0.4373137,0.05098967,0.0001071784],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.004866329,0.0006757962,0.9913248,0.000252945,0.0000581673,0.00007677974,0.0002024288,0.0005707425,0.001972032],"genre_scores_gemma":[0.06422652,0.00143605,0.9289873,0.0001723412,0.0002276217,0.0001489731,0.001442281,0.0002948702,0.003063974],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004942249,"threshold_uncertainty_score":0.01386499,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4400071817","doi":"10.1109/tkde.2024.3419698","title":"Laplacian Convolutional Representation for Traffic Time Series Imputation","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Traffic Prediction and Management Techniques","field":"Engineering","cited_by":31,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Polytechnique Montréal","funders":"Centre interuniversitaire de recherche sur les reseaux d'entreprise, la logistique et le transport","keywords":"Computer science; Time series; Imputation (statistics); Series (stratigraphy); Representation (politics); Data mining; Algorithm; Machine learning; Missing data; Geology","authors":[{"name":"Xinyu Chen","is_ca":true},{"name":"Zhanhong Cheng","is_ca":true},{"name":"HanQin Cai","is_ca":false},{"name":"Nicolas Saunier","is_ca":true},{"name":"Lijun Sun","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01660452875484313,"gpt":0.2639267242254463,"spread":0.2473221954706032,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001336544,0.0006583259,0.000734696,0.0009463005,0.0003790317,0.0006605837,0.001497766,0.001029202,0.001377517],"category_scores_gemma":[0.004696471,0.0003806003,0.000984519,0.001755424,0.000636645,0.001635437,0.0008282259,0.001688389,0.0004927144],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001171212,"about_ca_system_score_gemma":0.001710134,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01412958,"about_ca_topic_score_gemma":0.01136407,"domain_scores_codex":[0.9995671,0.0001107479,0.0000263267,0.0001227525,0.0001021604,0.00007089045],"domain_scores_gemma":[0.998825,0.0004984582,0.0001552647,0.0002140188,0.0002544876,0.00005273124],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000970746,0.00007383968,0.002134558,0.00006560802,0.00005842252,0.0001081331,0.00006612719,0.873587,0.002843123,0.03092026,0.003331183,0.0867146],"study_design_scores_gemma":[9.019355e-7,0.000002836671,0.00008361084,0.000001432683,0.000002724077,0.000005436908,0.000001865723,0.996882,0.0002092448,0.002657597,0.0001496888,0.000002679847],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02430127,0.0001850422,0.9734128,0.0003135455,0.00003027363,0.00001837627,0.0003013757,0.0006611149,0.0007762439],"genre_scores_gemma":[0.7967255,0.0007830239,0.1931181,0.0002478321,0.0001275255,0.0001432008,0.002232251,0.00019703,0.006425588],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01412958,"threshold_uncertainty_score":0.02809465,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1992440927","doi":"10.1109/tkde.2014.2359672","title":"Incorporating Social Role Theory into Topic Models for Social Media Content Analysis","year":2014,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Expert finding and Q&A systems","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"National Key Research and Development Program of China; Engineering and Physical Sciences Research Council; National Natural Science Foundation of China","keywords":"Social media; Computer science; Microblogging; Generative model; Focus (optics); Topic model; Regularization (linguistics); Data science; Process (computing); World Wide Web; Information retrieval; Generative grammar; Artificial intelligence","authors":[{"name":"Wayne Xin Zhao","is_ca":false},{"name":"Jinpeng Wang","is_ca":false},{"name":"Yulan He","is_ca":false},{"name":"Jian‐Yun Nie","is_ca":true},{"name":"Ji-Rong Wen","is_ca":false},{"name":"Xiaoming Li","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06034266008053929,"gpt":0.2760029146332962,"spread":0.2156602545527569,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006425377,0.001447709,0.001315838,0.004367538,0.0008014949,0.002552394,0.002628822,0.002170378,0.002415296],"category_scores_gemma":[0.0168967,0.0009832849,0.002768856,0.002945804,0.001736214,0.006138133,0.001593056,0.002619482,0.001433738],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002124449,"about_ca_system_score_gemma":0.001076591,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005579534,"about_ca_topic_score_gemma":0.006916485,"domain_scores_codex":[0.996286,0.002174425,0.0001638439,0.000683702,0.0004947983,0.0001971038],"domain_scores_gemma":[0.9878893,0.009994959,0.0006030026,0.0006701527,0.000587381,0.0002552312],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002781124,0.0003666921,0.01060478,0.0006882293,0.0005806728,0.0005680627,0.002945524,0.4143981,0.005604357,0.3608002,0.007779083,0.1953861],"study_design_scores_gemma":[0.00001855014,0.0000353814,0.0005890251,0.0000256431,0.00003903776,0.00007864011,0.000107673,0.9148447,0.0003721493,0.08084132,0.003024797,0.000023144],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008329822,0.0007646941,0.9886355,0.0005429049,0.00009435172,0.0001064318,0.00020897,0.0002436859,0.001073568],"genre_scores_gemma":[0.599721,0.002107651,0.3872825,0.0005045719,0.001148146,0.00111604,0.001444957,0.0003725703,0.006302607],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006425377,"threshold_uncertainty_score":0.03398108,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2109468254","doi":"10.1109/tkde.2012.86","title":"Achieving Data Privacy through Secrecy Views and Null-Based Virtual Updates","year":2012,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Logic, Reasoning, and Knowledge","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada; Technische Universität Wien","keywords":"Computer science; Null (SQL); Secrecy; Tuple; SQL; Semantics (computer science); Relational database; Information retrieval; Theoretical computer science; Database; Data mining; Programming language; Computer security; Mathematics","authors":[{"name":"Leopoldo Bertossi","is_ca":true},{"name":"Lechen Li","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06739210014020883,"gpt":0.3002273917608937,"spread":0.2328352916206849,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009718719,0.00083607,0.001011312,0.001406261,0.001749695,0.007988882,0.002574933,0.001401307,0.00158052],"category_scores_gemma":[0.01886802,0.0009864589,0.002204715,0.00176721,0.007563262,0.01994334,0.008944551,0.003789786,0.0005160403],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001458883,"about_ca_system_score_gemma":0.002978869,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001754887,"about_ca_topic_score_gemma":0.001491387,"domain_scores_codex":[0.9881644,0.004122072,0.00125864,0.001439778,0.003990895,0.00102419],"domain_scores_gemma":[0.9814098,0.006825322,0.001493824,0.007600015,0.00213544,0.0005356278],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001734016,0.00005596434,0.000852195,0.0000658881,0.00003699302,0.0001915758,0.0009607816,0.008203553,0.003223787,0.968259,0.001159872,0.01681685],"study_design_scores_gemma":[0.00006318793,0.00009548402,0.0001643835,0.00004316586,0.00008854787,0.000314434,0.0003222918,0.07279181,0.01576312,0.9010776,0.009224696,0.00005132245],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0313521,0.0002118035,0.9611979,0.001109178,0.00005680165,0.00009198327,0.0002114376,0.0007697699,0.004999137],"genre_scores_gemma":[0.7206561,0.0006648685,0.2711501,0.0008319452,0.0002600551,0.0002869899,0.0005196398,0.0004063641,0.00522398],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009718719,"threshold_uncertainty_score":0.0513981,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4396886508","doi":"10.1109/tkde.2024.3400824","title":"Natural Language Interfaces for Tabular Data Querying and Visualization: A Survey","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Visualization; Data visualization; Natural language; Information retrieval; Information visualization; Programming language; Database; World Wide Web; Natural language processing; Data mining","authors":[{"name":"Weixu Zhang","is_ca":false},{"name":"Yuanfeng Song","is_ca":false},{"name":"Victor Junqiu Wei","is_ca":false},{"name":"Yuxing Tian","is_ca":false},{"name":"Yiyan Qi","is_ca":false},{"name":"Jonathan H. Chan","is_ca":false},{"name":"Raymond Chi-Wing Wong","is_ca":false},{"name":"Haiqin Yang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04275557603702698,"gpt":0.3340007813895201,"spread":0.2912452053524931,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009524159,0.001648113,0.001857185,0.008704691,0.0007255728,0.007277838,0.003302865,0.001638042,0.01142746],"category_scores_gemma":[0.02837188,0.00109642,0.001651741,0.01007398,0.001394385,0.01492452,0.003029198,0.002554975,0.004918026],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001398369,"about_ca_system_score_gemma":0.001582293,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002229176,"about_ca_topic_score_gemma":0.00171096,"domain_scores_codex":[0.9929212,0.002145226,0.001085239,0.0006843958,0.002920064,0.0002438362],"domain_scores_gemma":[0.9692931,0.02429233,0.0008235521,0.00202702,0.003136843,0.0004272624],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002246576,0.0001864019,0.002273111,0.01087287,0.0001352516,0.0002108879,0.003087868,0.00153103,0.006245496,0.06207164,0.07334028,0.8398206],"study_design_scores_gemma":[0.00005092949,0.0001457877,0.002305418,0.004463352,0.0001073227,0.001351829,0.001281338,0.0128662,0.006602128,0.04475938,0.925866,0.0002003316],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.009992749,0.2623415,0.6530475,0.004802356,0.0005292224,0.0009412949,0.005453424,0.02872054,0.03417145],"genre_scores_gemma":[0.04739894,0.2179389,0.7016841,0.004003882,0.0008484941,0.001803239,0.01095466,0.005754135,0.009613626],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.01142746,"threshold_uncertainty_score":0.05036914,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2342411541","doi":"10.1109/tkde.2016.2527003","title":"Conflict-Aware Weighted Bipartite B-Matching and Its Application to E-Commerce","year":2016,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Optimization and Search Problems","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Victoria","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Bipartite graph; Computer science; Scalability; Matching (statistics); Scheduling (production processes); The Internet; Time complexity; Approximation algorithm; Blossom algorithm; Context (archaeology); Theoretical computer science; Data mining; Graph; Combinatorics; Algorithm; Mathematics; World Wide Web; Mathematical optimization; Database","authors":[{"name":"Chen Cheng","is_ca":true},{"name":"Lantian Zheng","is_ca":true},{"name":"Venkatesh Srinivasan","is_ca":true},{"name":"Alex Thomo","is_ca":true},{"name":"Kui Wu","is_ca":true},{"name":"Anthony Sukow","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03065758621392588,"gpt":0.2842775778446898,"spread":0.2536199916307639,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002932391,0.001156128,0.002013297,0.001890366,0.001763237,0.002008104,0.003116738,0.002614362,0.005286394],"category_scores_gemma":[0.01245274,0.0009676278,0.001644353,0.007187078,0.00133179,0.004582715,0.002490586,0.002471351,0.0009565239],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001682452,"about_ca_system_score_gemma":0.001831373,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005552815,"about_ca_topic_score_gemma":0.004563483,"domain_scores_codex":[0.997006,0.001409987,0.0001538893,0.0007262477,0.0004950973,0.0002087456],"domain_scores_gemma":[0.9953144,0.002899799,0.0004477366,0.0006876975,0.0004023817,0.0002480018],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003901254,0.0006668593,0.003358434,0.0007713311,0.0003151171,0.0003429276,0.0003187773,0.6191391,0.003694055,0.1496508,0.02207297,0.1992795],"study_design_scores_gemma":[0.0000459359,0.00005123621,0.0003640938,0.00002348473,0.00003324751,0.0001638708,0.00006614447,0.8667449,0.0006214534,0.1273107,0.004554823,0.00002024136],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04071235,0.002225564,0.9452356,0.001897225,0.0001587873,0.000289244,0.0007425092,0.0008764867,0.007862237],"genre_scores_gemma":[0.392518,0.001845643,0.5986553,0.0008003626,0.0002163891,0.0003670828,0.00160829,0.0003192657,0.003669688],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005552815,"threshold_uncertainty_score":0.01768476,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2147876569","doi":"10.1109/tkde.2015.2453171","title":"RankRC: Large-Scale Nonlinear Rare Class Ranking","year":2015,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Robustness (evolution); Kernel (algebra); Machine learning; Artificial intelligence; Focus (optics); Class (philosophy); Nonlinear system; Rare events; Algorithm; Computational complexity theory; Mathematics","authors":[{"name":"Aditya Tayal","is_ca":true},{"name":"Thomas F. Coleman","is_ca":true},{"name":"Yuying Li","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03825382451170466,"gpt":0.2748310372290403,"spread":0.2365772127173357,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004577007,0.001173652,0.00206592,0.001690308,0.001052807,0.002350374,0.003020836,0.002185794,0.003967816],"category_scores_gemma":[0.01490596,0.000438874,0.0008999616,0.001387311,0.001273424,0.00307397,0.002750654,0.002479194,0.002851157],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001093319,"about_ca_system_score_gemma":0.001906913,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003333721,"about_ca_topic_score_gemma":0.004138499,"domain_scores_codex":[0.9966751,0.001180379,0.0001613686,0.0005457434,0.001158357,0.0002790487],"domain_scores_gemma":[0.993258,0.002465226,0.0006489971,0.001930773,0.001405256,0.0002917683],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004838559,0.0004735461,0.002579386,0.0003290133,0.000146666,0.0002509505,0.0001757501,0.2211362,0.009610462,0.03754511,0.03168866,0.6955804],"study_design_scores_gemma":[0.00002588075,0.0001027312,0.0003540216,0.0000105435,0.00001170678,0.0001400602,0.00002966196,0.9795111,0.003635542,0.01377923,0.002375022,0.00002436884],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01646646,0.0003219786,0.9778265,0.0003190452,0.00008279479,0.0001498652,0.0002312261,0.003042616,0.001559528],"genre_scores_gemma":[0.3727706,0.0003319029,0.6167295,0.0004934108,0.0002044947,0.0003560472,0.00176154,0.0006580561,0.006694471],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004577007,"threshold_uncertainty_score":0.02420586,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2119387217","doi":"10.1109/tkde.2008.77","title":"Bias and Controversy in Evaluation Systems","year":2008,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Set (abstract data type); Object (grammar); Feature (linguistics); Information retrieval; Data set; Data science; Artificial intelligence; Data mining; Machine learning","authors":[{"name":"Hady W. Lauw","is_ca":false},{"name":"Ee‐Peng Lim","is_ca":false},{"name":"Ke Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09300923597913141,"gpt":0.2950924542262088,"spread":0.2020832182470774,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1517296,0.002222696,0.004257347,0.009435309,0.004118806,0.01283502,0.004737952,0.006168974,0.003401253],"category_scores_gemma":[0.4305668,0.001978989,0.001645238,0.005694316,0.01218211,0.01969589,0.009494485,0.005486071,0.0007946172],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.008358059,"about_ca_system_score_gemma":0.004304776,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002632966,"about_ca_topic_score_gemma":0.001340357,"domain_scores_codex":[0.7619199,0.1573387,0.01401956,0.02521574,0.03882136,0.002684615],"domain_scores_gemma":[0.5184613,0.3986451,0.02932931,0.02144172,0.02894537,0.003177179],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006813184,0.0002081588,0.02717591,0.001359739,0.0007981609,0.0004464717,0.004807076,0.06320614,0.001719681,0.6151987,0.005322352,0.2790763],"study_design_scores_gemma":[0.0002171686,0.000265448,0.004760244,0.0004300561,0.000268768,0.0005529802,0.0005120935,0.1973519,0.002101582,0.7798671,0.01347438,0.0001982745],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04278415,0.006982482,0.9258305,0.007486515,0.0004036028,0.000756843,0.00019562,0.0007079719,0.01485234],"genre_scores_gemma":[0.7814409,0.001385486,0.2112347,0.001368045,0.0009421436,0.0008920407,0.0001956064,0.0001639573,0.00237699],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.1517296,"threshold_uncertainty_score":0.8024325,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2766474394","doi":"10.1109/tkde.2017.2767044","title":"Workload Management in Database Management Systems: A Taxonomy","year":2017,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Cloud Computing and Resource Management","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Queen's University","funders":"","keywords":"Workload; Computer science; Database; Process (computing); Data management; Operating system","authors":[{"name":"Mingyi Zhang","is_ca":false},{"name":"Patrick Martin","is_ca":true},{"name":"Wendy Powley","is_ca":true},{"name":"Jianjun Chen","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03969699418589224,"gpt":0.2605875175153202,"spread":0.2208905233294279,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005687536,0.001542475,0.001650756,0.01059446,0.002836982,0.008733058,0.004042876,0.003651532,0.001219476],"category_scores_gemma":[0.01112736,0.001330882,0.001604234,0.01492093,0.002492314,0.01292624,0.002868394,0.003275684,0.0008359441],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00307061,"about_ca_system_score_gemma":0.004090284,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003525076,"about_ca_topic_score_gemma":0.002504327,"domain_scores_codex":[0.9892837,0.001707427,0.0023166,0.001200401,0.004766142,0.0007257919],"domain_scores_gemma":[0.9873591,0.004714068,0.00207778,0.001254171,0.003818275,0.0007766061],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002124961,0.00081983,0.02533307,0.006762943,0.0002205683,0.0007196634,0.005658437,0.01701293,0.009630371,0.1940921,0.02164022,0.7178974],"study_design_scores_gemma":[0.0001072773,0.001023903,0.02525958,0.00702953,0.000278403,0.00685819,0.006641834,0.1394917,0.006541251,0.2740376,0.5321996,0.0005311456],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06177786,0.215094,0.6730722,0.008962342,0.001242346,0.002990784,0.001286246,0.002469243,0.03310495],"genre_scores_gemma":[0.1592087,0.1055401,0.7208831,0.00201656,0.00156514,0.001834702,0.002665975,0.0002914667,0.005994361],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01059446,"threshold_uncertainty_score":0.03007895,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2051664277","doi":"10.1109/tkde.2011.100","title":"Discovery of Delta Closed Patterns and Noninduced Patterns from Sequences","year":2011,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Substring; Computer science; Set (abstract data type); Sequence (biology); Data mining; Suffix tree; Pattern recognition (psychology); Artificial intelligence; Data structure","authors":[{"name":"Andrew K. C. Wong","is_ca":true},{"name":"Dennis Zhuang","is_ca":true},{"name":"Gary C.L. Li","is_ca":true},{"name":"En-Shiun Annie Lee","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04450856788129964,"gpt":0.2576284759502814,"spread":0.2131199080689817,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001274288,0.0004399205,0.0008899202,0.002038661,0.0004573892,0.001088422,0.0009321105,0.0007588182,0.0008824806],"category_scores_gemma":[0.007479352,0.0003381285,0.0008301043,0.002100454,0.0006206754,0.001937711,0.0009335481,0.000871457,0.0004542625],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002005755,"about_ca_system_score_gemma":0.0006847173,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002502097,"about_ca_topic_score_gemma":0.0003905424,"domain_scores_codex":[0.9988662,0.0002290672,0.0001949036,0.0002354001,0.0004031723,0.0000712792],"domain_scores_gemma":[0.9947078,0.0028836,0.0007511523,0.0008035665,0.0006859981,0.0001678541],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007537596,0.0004616674,0.03196485,0.0008488065,0.000165526,0.001485954,0.0007645803,0.02285811,0.07069928,0.02014047,0.001374949,0.8484821],"study_design_scores_gemma":[0.0001306563,0.0008147806,0.01873796,0.0001856492,0.0001424196,0.004656313,0.0009085353,0.763554,0.0814157,0.1200418,0.009317835,0.00009429782],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2312397,0.0005761514,0.7655904,0.0002042731,0.0000327663,0.0001497367,0.0004342442,0.0006380317,0.001134775],"genre_scores_gemma":[0.401284,0.0004770832,0.5948402,0.00009234896,0.00003297558,0.0001380651,0.002047975,0.00007593528,0.001011468],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002038661,"threshold_uncertainty_score":0.00673914,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1978421404","doi":"10.1109/tkde.2014.3","title":"Editorial [State of the Transactions]","year":2013,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Service-Oriented Architecture and Web Services","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Dalhousie University; University of Ottawa","keywords":"Computer science; Library science; Reputation; Operations research; Political science; Mathematics; Law","authors":[{"name":"Jian Pei","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.008794747565106346,"gpt":0.217637647023636,"spread":0.2088428994585297,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005740683,0.00261817,0.003088774,0.004619345,0.002966447,0.01460079,0.00410713,0.006659096,0.1363899],"category_scores_gemma":[0.03675474,0.001119309,0.002719971,0.002645719,0.002256368,0.005773509,0.002319718,0.009628485,0.1585513],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002568043,"about_ca_system_score_gemma":0.00581351,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001088205,"about_ca_topic_score_gemma":0.00180623,"domain_scores_codex":[0.9909585,0.0007519232,0.001264409,0.001205121,0.005142122,0.0006780087],"domain_scores_gemma":[0.9501776,0.00392235,0.003144039,0.002029388,0.03325886,0.007467686],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002360807,0.000007667663,0.00002022304,0.0002093597,0.000008189631,0.00005328107,0.0000067816,0.00001852609,0.00007147958,0.000184855,0.9855816,0.01381444],"study_design_scores_gemma":[0.00001789401,0.0000183096,0.00009513114,0.0002940545,0.000009903792,0.0001195547,0.00001716033,0.00004920399,0.0000808949,0.000312422,0.9989723,0.00001312594],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.00004977302,0.003020404,0.0002732871,0.01149066,0.9778638,0.00005412299,0.0001893934,0.0002710437,0.006787477],"genre_scores_gemma":[0.0008713854,0.008478979,0.000546437,0.01666776,0.9133981,0.00009546692,0.0005442301,0.0003717473,0.0590259],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.1363899,"threshold_uncertainty_score":0.4562697,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2114175725","doi":"10.1109/tkde.2009.49","title":"Enhancing Learning Objects with an Ontology-Based Memory","year":2009,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Computer science; Reusability; Learning object; Ontology; Semantic Web; Process (computing); Knowledge base; Artificial intelligence; Natural language processing; World Wide Web; Programming language","authors":[{"name":"Amal Zouaq","is_ca":true},{"name":"Roger Nkambou","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01977965045427822,"gpt":0.2566786341254655,"spread":0.2368989836711873,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001097899,0.0003121765,0.0003860597,0.0009362364,0.0007836079,0.002978392,0.001763621,0.000814438,0.00201338],"category_scores_gemma":[0.003409038,0.000340399,0.0006977285,0.0009684542,0.001112577,0.008829594,0.003842257,0.001094623,0.0008101422],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006360203,"about_ca_system_score_gemma":0.001371662,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003223082,"about_ca_topic_score_gemma":0.003901459,"domain_scores_codex":[0.9994169,0.00009352279,0.00008715405,0.0001069983,0.0002149697,0.00008044052],"domain_scores_gemma":[0.9985123,0.0003040167,0.0001597756,0.0006459436,0.0002441667,0.0001338072],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002186104,0.0007561985,0.004561213,0.000505617,0.0001273068,0.0005880402,0.003742196,0.01888233,0.0357227,0.2806393,0.008167138,0.6460894],"study_design_scores_gemma":[0.0001758751,0.0005053799,0.003597846,0.0003339277,0.0006041164,0.00158077,0.002884198,0.2454037,0.08920144,0.3155843,0.3399311,0.0001975169],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08648232,0.000659201,0.8813296,0.001824863,0.0001912738,0.0002507865,0.0001924516,0.003725019,0.0253445],"genre_scores_gemma":[0.3993351,0.0007762984,0.5864334,0.0004033098,0.00007622405,0.0002337238,0.0004802952,0.0002413111,0.01202035],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003223082,"threshold_uncertainty_score":0.006735444,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2152225808","doi":"10.1109/tkde.2010.33","title":"Asking Generalized Queries to Domain Experts to Improve Learning","year":2010,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Machine Learning and Algorithms","field":"Computer Science","cited_by":25,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"China University of Geosciences; Shanghai Jiao Tong University; University of Pennsylvania","keywords":"Computer science; Oracle; Construct (python library); Classifier (UML); Ask price; Domain (mathematical analysis); Set (abstract data type); Machine learning; Class (philosophy); Labeled data; Active learning (machine learning); Artificial intelligence; Theoretical computer science; Mathematics","authors":[{"name":"Jun Du","is_ca":true},{"name":"Charles X. Ling","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01083209693741239,"gpt":0.265796040034446,"spread":0.2549639430970336,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00988446,0.002244467,0.003020145,0.001613018,0.001090073,0.002825211,0.004128054,0.003694931,0.003855173],"category_scores_gemma":[0.04832206,0.0008146566,0.001162613,0.001757933,0.001898551,0.0093127,0.00430238,0.004149685,0.001297158],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001454705,"about_ca_system_score_gemma":0.001849134,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001917717,"about_ca_topic_score_gemma":0.003061764,"domain_scores_codex":[0.9883876,0.00538301,0.0008590032,0.002462897,0.002352913,0.0005544581],"domain_scores_gemma":[0.9429019,0.04266129,0.002257681,0.007372573,0.003924748,0.0008816862],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001373935,0.001426422,0.01162209,0.0005643048,0.0002502292,0.0003642141,0.001998744,0.2347496,0.01768917,0.03109579,0.01708812,0.6817774],"study_design_scores_gemma":[0.0001074122,0.0001735658,0.0005102205,0.00001762523,0.00003922844,0.0001191885,0.0002555104,0.9607071,0.005059374,0.03015342,0.002831169,0.00002626151],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04329183,0.0003066156,0.9513758,0.0007928361,0.00004704852,0.0002276499,0.0001200994,0.002371843,0.001466188],"genre_scores_gemma":[0.5127568,0.0002067417,0.4824381,0.0008016307,0.000201756,0.0004497688,0.0008665992,0.0003196863,0.001958931],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00988446,"threshold_uncertainty_score":0.05227464,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}