{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":49,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":49,"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":"ad6497fcacec","filters":{"venue":"Data Mining and Knowledge Discovery"}},"results":[{"id":"W2115482638","doi":"10.1023/b:dami.0000005258.31418.83","title":"Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach","year":2003,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":2615,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Data mining; Computer science; Scalability; Tree (set theory); Set (abstract data type); Apriori algorithm; GSP Algorithm; Association rule learning; Trie; Tree structure; Pattern recognition (psychology); Data structure; Artificial intelligence; Database; Mathematics","authors":[{"name":"Jiawei Han","is_ca":false},{"name":"Jian Pei","is_ca":false},{"name":"Yiwen Yin","is_ca":true},{"name":"Runying Mao","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06223377143017436,"gpt":0.2906354910060366,"spread":0.2284017195758623,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005180415,0.001300011,0.003232713,0.004975306,0.001423573,0.003417117,0.00434752,0.002591436,0.002064071],"category_scores_gemma":[0.02203915,0.001061476,0.00262189,0.006072121,0.001022005,0.006168364,0.001920203,0.002250755,0.001881533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004025765,"about_ca_system_score_gemma":0.001824103,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001007755,"about_ca_topic_score_gemma":0.001539642,"domain_scores_codex":[0.9948336,0.001288063,0.0005897361,0.001057398,0.001971786,0.0002594239],"domain_scores_gemma":[0.9828394,0.01182238,0.0009671408,0.001924336,0.002108966,0.0003377635],"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":[0.001031068,0.001224696,0.01331,0.001352846,0.0007791875,0.002112762,0.0004719143,0.03880866,0.01247221,0.02362365,0.008990197,0.8958228],"study_design_scores_gemma":[0.0002655924,0.0006092436,0.002345708,0.0001688598,0.0007396905,0.003678299,0.0003230342,0.8846376,0.008669051,0.0908212,0.007629762,0.000112],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01554049,0.0007677517,0.9811126,0.0003627418,0.0000735874,0.0002575588,0.0005567197,0.0007541402,0.0005744072],"genre_scores_gemma":[0.09764457,0.0006919973,0.8979678,0.0002048634,0.0001523225,0.000236876,0.001938123,0.0001178523,0.001045749],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005180415,"threshold_uncertainty_score":0.02739698,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2282861635","doi":"10.1007/s10618-015-0444-8","title":"On the evaluation of unsupervised outlier detection: measures, datasets, and an empirical study","year":2016,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":732,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada; Fundação de Amparo à Pesquisa do Estado de São Paulo; Teknologi og Produktion, Det Frie Forskningsråd","keywords":"Outlier; Anomaly detection; Benchmark (surveying); Computer science; Data mining; Artificial intelligence; Ground truth; Pattern recognition (psychology); Set (abstract data type); Machine learning","authors":[{"name":"Guilherme O. Campos","is_ca":false},{"name":"Arthur Zimek","is_ca":false},{"name":"Jörg Sander","is_ca":true},{"name":"Ricardo J. G. B. Campello","is_ca":false},{"name":"Barbora Micenková","is_ca":false},{"name":"Erich Schubert","is_ca":false},{"name":"Ira Assent","is_ca":false},{"name":"Michael E. Houle","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1553202106037584,"gpt":0.3896393330565236,"spread":0.2343191224527651,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03498443,0.001710232,0.001651135,0.00700916,0.001441702,0.003051224,0.00281065,0.002473269,0.0006181667],"category_scores_gemma":[0.1873556,0.000367881,0.001387332,0.005792591,0.00221568,0.005608084,0.002598465,0.001759048,0.0002410891],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002109847,"about_ca_system_score_gemma":0.001821054,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004075689,"about_ca_topic_score_gemma":0.004933161,"domain_scores_codex":[0.9454237,0.02698848,0.006062921,0.005060516,0.01550229,0.00096213],"domain_scores_gemma":[0.658196,0.2746069,0.01788623,0.02211152,0.02525189,0.001947556],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.006365941,0.004265401,0.3466473,0.002784813,0.003256323,0.000417447,0.001608432,0.189528,0.009163713,0.01165585,0.01500079,0.409306],"study_design_scores_gemma":[0.0003832132,0.004315083,0.07509547,0.0003935689,0.0008865569,0.001528291,0.001402172,0.8831693,0.01294639,0.01522497,0.004442294,0.0002127602],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7978086,0.005405053,0.1867845,0.0009972693,0.0003056825,0.0004534209,0.00341261,0.001684273,0.003148376],"genre_scores_gemma":[0.9221168,0.0006867843,0.070075,0.00014896,0.0001161745,0.0001700518,0.00605825,0.0001733945,0.000454681],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9650156,"threshold_uncertainty_score":0.1850175,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2075949491","doi":"10.1007/s10618-012-0300-z","title":"Local outlier detection reconsidered: a generalized view on locality with applications to spatial, video, and network outlier detection","year":2012,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":294,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Locality; Anomaly detection; Computer science; Outlier; Generalization; Data mining; Artificial intelligence; Graph; Machine learning; Pattern recognition (psychology); Theoretical computer science; Mathematics","authors":[{"name":"Erich Schubert","is_ca":false},{"name":"Arthur Zimek","is_ca":true},{"name":"Hans‐Peter Kriegel","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04074681739410198,"gpt":0.2935640976154808,"spread":0.2528172802213788,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007113142,0.001156019,0.004196495,0.005036319,0.002090828,0.008104287,0.005604001,0.0041167,0.001522881],"category_scores_gemma":[0.0289917,0.0008857927,0.002795294,0.007148171,0.009988977,0.01538311,0.00696527,0.006032175,0.0003173563],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002001858,"about_ca_system_score_gemma":0.001991178,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006724066,"about_ca_topic_score_gemma":0.005623489,"domain_scores_codex":[0.992952,0.002359,0.0005071025,0.001803861,0.001924208,0.0004537677],"domain_scores_gemma":[0.970076,0.01657313,0.002522551,0.004989086,0.004783046,0.001056188],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001943456,0.0001674396,0.01176181,0.0004647663,0.000517266,0.001004048,0.001061243,0.1191879,0.002499627,0.7001902,0.006435274,0.1565161],"study_design_scores_gemma":[0.00002378635,0.00008779946,0.00112533,0.00006865357,0.0001310337,0.0003778785,0.0003321755,0.3962066,0.000793275,0.5962981,0.004484469,0.00007100717],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.012785,0.002938383,0.9758567,0.005835937,0.000201621,0.00003135544,0.0001120204,0.0001915215,0.002047398],"genre_scores_gemma":[0.6994204,0.005672949,0.2863863,0.001773779,0.003280421,0.0001521121,0.0002508268,0.0002075071,0.002855724],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008104287,"threshold_uncertainty_score":0.03761834,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1972588194","doi":"10.1007/s10618-009-0156-z","title":"Time to CARE: a collaborative engine for practical disease prediction","year":2009,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":147,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Centers for Disease Control and Prevention; McGill University","keywords":"Computer science; Action (physics); Disease; Collaborative filtering; Health care; Order (exchange); Risk analysis (engineering); Medical costs; Machine learning; Recommender system; Medicine","authors":[{"name":"Darcy Davis","is_ca":false},{"name":"Nitesh V. Chawla","is_ca":false},{"name":"Nicholas A. Christakis","is_ca":false},{"name":"Albert-Ĺaszló Barabási","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03094374567114801,"gpt":0.3502158157660921,"spread":0.3192720700949441,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004550647,0.001696787,0.001599719,0.00499421,0.001018927,0.002441066,0.002595331,0.001816564,0.01270359],"category_scores_gemma":[0.02141987,0.0006368564,0.00170615,0.003975022,0.0003652547,0.003487008,0.003120387,0.001243705,0.005183405],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007574023,"about_ca_system_score_gemma":0.002201877,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01222419,"about_ca_topic_score_gemma":0.01446168,"domain_scores_codex":[0.997992,0.0003554139,0.0003705518,0.0004445064,0.0007331418,0.0001043171],"domain_scores_gemma":[0.9897804,0.005486603,0.0005616357,0.002138265,0.001198175,0.0008348265],"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":[0.004516285,0.0007943138,0.02490761,0.001094922,0.0009265026,0.001082815,0.001129456,0.02632057,0.006268725,0.02497943,0.1950352,0.7129441],"study_design_scores_gemma":[0.0007777857,0.0004776991,0.006020411,0.0002442883,0.001073979,0.001141295,0.000392621,0.7235826,0.02347932,0.08046527,0.1620131,0.0003315003],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02557091,0.001487862,0.8023174,0.002100249,0.0007694867,0.0009059707,0.02503499,0.1328095,0.009003695],"genre_scores_gemma":[0.1966566,0.001185873,0.7441306,0.00121633,0.0003565917,0.000909315,0.0412139,0.004131437,0.01019937],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01270359,"threshold_uncertainty_score":0.04249775,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2078064242","doi":"10.1023/b:dami.0000023674.74932.4c","title":"Pushing Convertible Constraints in Frequent Itemset Mining","year":2004,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":134,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Constraint (computer-aided design); Computer science; Data mining; Context (archaeology); Association rule learning; Convertible; A priori and a posteriori; Class (philosophy); Apriori algorithm; Monotonic function; Sequential Pattern Mining; Artificial intelligence; Mathematics; Engineering; Geography","authors":[{"name":"Jian Pei","is_ca":false},{"name":"Jiawei Han","is_ca":false},{"name":"Laks V. S. Lakshmanan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04812423122534112,"gpt":0.2999252304469889,"spread":0.2518009992216478,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007770571,0.001120368,0.002495039,0.005193333,0.001746164,0.005434884,0.004381738,0.002693982,0.005785996],"category_scores_gemma":[0.07125326,0.001964246,0.002228376,0.007130831,0.002327159,0.01350291,0.005022739,0.005739216,0.001191658],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000814135,"about_ca_system_score_gemma":0.001601174,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003388511,"about_ca_topic_score_gemma":0.003720558,"domain_scores_codex":[0.9913293,0.002363333,0.001001851,0.001404176,0.003482599,0.0004185825],"domain_scores_gemma":[0.9512032,0.03725308,0.002019739,0.005807577,0.003012704,0.0007036111],"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":[0.00125527,0.0006220003,0.009343117,0.001632052,0.0004388645,0.003067011,0.001541019,0.1079403,0.004943031,0.2536542,0.01321659,0.6023466],"study_design_scores_gemma":[0.000134146,0.0002046629,0.001323905,0.0003445988,0.0001979504,0.001268606,0.0004854435,0.428769,0.004620282,0.5465281,0.01603468,0.00008863499],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06120999,0.003186459,0.9239631,0.002204013,0.0004418405,0.0003971095,0.001540459,0.00144744,0.005609586],"genre_scores_gemma":[0.3256478,0.002244064,0.6614846,0.000805379,0.0004445596,0.0003639216,0.003477396,0.0003387751,0.005193418],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007770571,"threshold_uncertainty_score":0.0410952,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2116064560","doi":"10.1007/s10618-005-0248-3","title":"Mining Web Log Sequential Patterns with Position Coded Pre-Order Linked WAP-Tree","year":2005,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":131,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"","keywords":"Computer science; Web log analysis software; Tree (set theory); Trie; Suffix tree; Prefix; Data mining; Subsequence; Web mining; Node (physics); Web server; Fractal tree index; Tree structure; Interval tree; Web page; Data structure; Algorithm; The Internet; World Wide Web; Mathematics; Binary tree; Static web page","authors":[{"name":"C. I. Ezeife","is_ca":true},{"name":"Yi Lu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03043223887843423,"gpt":0.2874098398765441,"spread":0.2569776009981099,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005223575,0.000408365,0.0005828464,0.004766338,0.0008250256,0.001168149,0.0009660025,0.0007547274,0.001963287],"category_scores_gemma":[0.005336094,0.0003361931,0.0007076811,0.004870439,0.0003288419,0.001636212,0.0006796703,0.0007584738,0.001058037],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004136906,"about_ca_system_score_gemma":0.001311161,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003828702,"about_ca_topic_score_gemma":0.008566475,"domain_scores_codex":[0.9993704,0.00007417148,0.00007431507,0.0001242522,0.0002737573,0.00008307207],"domain_scores_gemma":[0.997134,0.001107086,0.0003126317,0.0005005373,0.0007855956,0.0001601738],"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":[0.001794458,0.001291771,0.1252399,0.0007896124,0.0002964315,0.002702913,0.0005911925,0.07023752,0.02773121,0.01470895,0.01710113,0.737515],"study_design_scores_gemma":[0.00007652842,0.0003251839,0.01447759,0.00008038871,0.0001469751,0.001153518,0.000408414,0.93617,0.01384274,0.02474358,0.008537252,0.00003790231],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5177175,0.0005613393,0.4576658,0.0004768656,0.0001407863,0.0006583247,0.01054572,0.007308394,0.004925318],"genre_scores_gemma":[0.7005172,0.0001977091,0.2846425,0.00008240846,0.00004467002,0.0002708926,0.01110097,0.0001843702,0.002959259],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.004766338,"threshold_uncertainty_score":0.007612824,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3141059875","doi":"10.1007/s10618-005-1355-x","title":"Mining Customer Value: From Association Rules to Direct Marketing","year":2005,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Consumer Market Behavior and Pricing","field":"Business, Management and Accounting","cited_by":117,"is_retracted":false,"has_abstract":false,"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","keywords":"Association rule learning; Profit (economics); Computer science; Data mining; Liberian dollar; Direct marketing; Customer lifetime value; Customer value; Marketing; Artificial intelligence; Machine learning; Customer retention; Business; Economics; Finance","authors":[{"name":"Ke Wang Wong","is_ca":true},{"name":"Senqiang Zhou","is_ca":true},{"name":"Qiang Yang","is_ca":false},{"name":"Jack Man Shun Yeung","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03081130680570864,"gpt":0.2724500237614445,"spread":0.2416387169557359,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002088039,0.001116155,0.001299176,0.008165473,0.0007216602,0.00251613,0.001782368,0.001437195,0.001282103],"category_scores_gemma":[0.01577572,0.0005571874,0.001514751,0.00614301,0.0006588196,0.002311364,0.001033445,0.001928946,0.001094354],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005341206,"about_ca_system_score_gemma":0.001046857,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004753905,"about_ca_topic_score_gemma":0.005938932,"domain_scores_codex":[0.9979576,0.0005599029,0.0002079485,0.0004626282,0.0006711259,0.0001408293],"domain_scores_gemma":[0.9871471,0.01054532,0.0006706488,0.0006104533,0.0007965648,0.0002300155],"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":[0.0008967638,0.002187593,0.1941845,0.0005765673,0.001283958,0.001514114,0.0005800849,0.03053873,0.003435023,0.005766421,0.01333312,0.7457032],"study_design_scores_gemma":[0.0002877536,0.000522409,0.06043594,0.000324012,0.001650183,0.002157105,0.001089938,0.8033013,0.008595553,0.1095497,0.01197229,0.0001137416],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7287794,0.00789167,0.2405359,0.003062147,0.0003329052,0.0004379136,0.008463337,0.002633769,0.007862872],"genre_scores_gemma":[0.8699178,0.001703076,0.1156029,0.0003999568,0.0003359499,0.0002374785,0.009897125,0.0000994978,0.001806039],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.008165473,"threshold_uncertainty_score":0.01104271,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3043874967","doi":"10.1007/s10618-020-00705-9","title":"Deep soccer analytics: learning an action-value function for evaluating soccer players","year":2020,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":104,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Artificial intelligence; Computer science; Analytics; Value (mathematics); Action (physics); Function (biology); Machine learning; Data mining","authors":[{"name":"Guiliang Liu","is_ca":true},{"name":"Yudong Luo","is_ca":true},{"name":"Oliver Schulte","is_ca":true},{"name":"Tarak Kharrat","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2560302781838675,"gpt":0.3534881577537748,"spread":0.09745787956990726,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001120276,0.001102587,0.0008141145,0.0012437,0.000267885,0.001074301,0.001063214,0.001127955,0.002993701],"category_scores_gemma":[0.003671549,0.0002986868,0.000556568,0.0009973476,0.0004183701,0.001267576,0.001179066,0.00146882,0.000745171],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008873035,"about_ca_system_score_gemma":0.001374188,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009037605,"about_ca_topic_score_gemma":0.01287166,"domain_scores_codex":[0.9995926,0.0001130073,0.00002168332,0.0001214315,0.00007229386,0.00007898398],"domain_scores_gemma":[0.9990242,0.0005040431,0.0001185361,0.00007955667,0.0001531973,0.0001206203],"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":[0.001059369,0.001203737,0.05778038,0.0001777316,0.000330985,0.0001045245,0.0001459815,0.3711831,0.003625823,0.010875,0.01353667,0.5399767],"study_design_scores_gemma":[0.00001376481,0.00006922209,0.003177723,0.00001681296,0.00001673423,0.00001303527,0.00002169622,0.9846789,0.0007260299,0.01083131,0.0004263553,0.000008396543],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3714152,0.001472507,0.6124872,0.001993577,0.0001835553,0.0001878274,0.004094609,0.002615131,0.00555042],"genre_scores_gemma":[0.9575452,0.0002926756,0.0374999,0.0001345062,0.00006437733,0.00008704273,0.001894221,0.00004598638,0.002435997],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009037605,"threshold_uncertainty_score":0.01797003,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2159328231","doi":"10.1007/s10618-007-0083-9","title":"Mining functional dependencies from data","year":2007,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":90,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Regina","funders":"","keywords":"Correctness; Pruning; Computer science; Data mining; Information loss; Process (computing); Functional dependency; Series (stratigraphy); Axiom; Algorithm; Artificial intelligence; Mathematics; Relational database","authors":[{"name":"Hong Yao","is_ca":true},{"name":"Howard J. Hamilton","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1034148320734652,"gpt":0.31561346603797,"spread":0.2121986339645047,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001807192,0.001145129,0.0008195778,0.006528472,0.001175276,0.001336685,0.001511994,0.0008547136,0.003158364],"category_scores_gemma":[0.01611426,0.0007398302,0.001586476,0.003767151,0.0007760127,0.003915151,0.001456552,0.001728425,0.001375173],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006388622,"about_ca_system_score_gemma":0.001992023,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002873915,"about_ca_topic_score_gemma":0.00632355,"domain_scores_codex":[0.9977489,0.0003575693,0.0002097216,0.0005019616,0.001018464,0.0001633268],"domain_scores_gemma":[0.9883692,0.007944746,0.0008204216,0.001281746,0.001312535,0.0002712941],"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":[0.0007091836,0.0007534022,0.0774849,0.002092498,0.0006063586,0.004297322,0.001099995,0.03430695,0.02455125,0.05972816,0.03387256,0.7604974],"study_design_scores_gemma":[0.000118062,0.0004775107,0.03619569,0.0007064652,0.001148407,0.004627888,0.001053392,0.5416309,0.04005851,0.3049492,0.06888306,0.0001510532],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2499745,0.003117688,0.7079066,0.002600701,0.0002769143,0.0006900608,0.01891971,0.006621375,0.009892497],"genre_scores_gemma":[0.6248855,0.00204165,0.3247882,0.000416424,0.0002497997,0.0006180092,0.04115873,0.0006147268,0.005227012],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006528472,"threshold_uncertainty_score":0.01056576,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2079040080","doi":"10.1007/s10618-013-0311-4","title":"A framework for semi-supervised and unsupervised optimal extraction of clusters from hierarchies","year":2013,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":83,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Cluster analysis; Computer science; Tree (set theory); Cluster (spacecraft); Variety (cybernetics); Data mining; Artificial intelligence; Extraction (chemistry); Machine learning; Mathematics","authors":[{"name":"Ricardo J. G. B. Campello","is_ca":true},{"name":"Davoud Moulavi","is_ca":true},{"name":"Arthur Zimek","is_ca":true},{"name":"Jörg Sander","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06040773208466434,"gpt":0.3443362713711306,"spread":0.2839285392864663,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004712904,0.00117388,0.002136,0.003472786,0.001937464,0.003710632,0.004754857,0.001975002,0.002734884],"category_scores_gemma":[0.01164817,0.001068743,0.002507082,0.004050041,0.002121405,0.003801972,0.004615476,0.002957153,0.001830198],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001660069,"about_ca_system_score_gemma":0.003871193,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008027991,"about_ca_topic_score_gemma":0.01518997,"domain_scores_codex":[0.9961482,0.001162313,0.0003094972,0.0008732014,0.001257979,0.0002488778],"domain_scores_gemma":[0.9951792,0.001775514,0.0003824513,0.001027828,0.001393923,0.0002410205],"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":[0.0002108904,0.0002091988,0.001290277,0.0005154728,0.0002858337,0.0002536786,0.0008870711,0.1577751,0.01065595,0.2780117,0.01594564,0.5339591],"study_design_scores_gemma":[0.00003342421,0.00004685997,0.0003688262,0.00006355796,0.00005094475,0.0001592612,0.0001190532,0.7681822,0.002756648,0.2180263,0.01014296,0.00005003229],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0003578806,0.00006508255,0.999009,0.00004375724,0.000009231741,0.00002323801,0.0000526012,0.0002874676,0.0001516203],"genre_scores_gemma":[0.01760601,0.0001191692,0.9808699,0.00005842078,0.00004026802,0.0001329232,0.0003835343,0.0001414221,0.0006483616],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008027991,"threshold_uncertainty_score":0.02492452,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1603563923","doi":"10.1023/a:1022419032620","title":"Extracting Share Frequent Itemsets with Infrequent Subsets","year":2003,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":77,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Regina","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Data mining; Association rule learning; Computer science; Measure (data warehouse); Property (philosophy); Database transaction; Heuristic; Set (abstract data type); Closure (psychology); Artificial intelligence; Database","authors":[{"name":"Brock Barber","is_ca":true},{"name":"Howard J. Hamilton","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.051266264070474,"gpt":0.2920698709287232,"spread":0.2408036068582492,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001572488,0.001363075,0.002251542,0.007360559,0.001557481,0.002753854,0.001869779,0.001460448,0.003579656],"category_scores_gemma":[0.01044916,0.00090428,0.002217067,0.007606711,0.0005187829,0.003946573,0.001948068,0.001152782,0.002578871],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003702138,"about_ca_system_score_gemma":0.001306992,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007377396,"about_ca_topic_score_gemma":0.001458865,"domain_scores_codex":[0.9976024,0.0002736527,0.0003859111,0.0004372137,0.001030043,0.0002708836],"domain_scores_gemma":[0.9924135,0.002974762,0.0008845992,0.001658465,0.00168992,0.0003788809],"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":[0.003026271,0.001073537,0.07644452,0.0009057978,0.001165574,0.005128957,0.001076589,0.01001131,0.05625634,0.008061985,0.0090631,0.827786],"study_design_scores_gemma":[0.000765398,0.003734115,0.07774067,0.0004411373,0.003893983,0.02375501,0.005566013,0.6074957,0.1432663,0.08531866,0.04776016,0.000262851],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4721496,0.001146708,0.5150251,0.0005561928,0.0003352791,0.0007955026,0.003467538,0.002309663,0.004214488],"genre_scores_gemma":[0.6072561,0.000567103,0.3765204,0.0001180013,0.0004167677,0.000305962,0.011917,0.000196841,0.002701821],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007360559,"threshold_uncertainty_score":0.01197511,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2086829791","doi":"10.1007/s10618-011-0221-2","title":"DHCC: Divisive hierarchical clustering of categorical data","year":2011,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":77,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Categorical variable; Cluster analysis; Computer science; Data mining; Initialization; Similarity (geometry); Linear subspace; Single-linkage clustering; Hierarchical clustering; Artificial intelligence; Pattern recognition (psychology); CURE data clustering algorithm; Correlation clustering; Machine learning; Mathematics","authors":[{"name":"Tengke Xiong","is_ca":true},{"name":"Shengrui Wang","is_ca":true},{"name":"André Mayers","is_ca":true},{"name":"Ernest Monga","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1924145629485619,"gpt":0.3565834104161071,"spread":0.1641688474675453,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003186143,0.001350303,0.001641656,0.005162863,0.001626792,0.002618825,0.003455608,0.001392419,0.006119874],"category_scores_gemma":[0.01450486,0.0008620228,0.00202511,0.007430254,0.0006719409,0.001627894,0.00278945,0.002248004,0.004251963],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001201059,"about_ca_system_score_gemma":0.003998428,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01226791,"about_ca_topic_score_gemma":0.01728992,"domain_scores_codex":[0.9966754,0.001123483,0.000236967,0.0005648898,0.001192503,0.0002068261],"domain_scores_gemma":[0.9951391,0.001474113,0.0002268712,0.00158372,0.001367019,0.0002092161],"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":[0.000424089,0.0002477535,0.003628014,0.0007672905,0.0004126163,0.0001602235,0.000566676,0.04784878,0.005984741,0.02416062,0.06596284,0.8498363],"study_design_scores_gemma":[0.0001696962,0.0001659042,0.003628456,0.000129551,0.0001926012,0.0003882017,0.0002982659,0.839159,0.0140065,0.07377815,0.06792869,0.0001550078],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003337046,0.0003346055,0.9810245,0.0001645491,0.0001614567,0.0004255908,0.002433635,0.01115917,0.0009594607],"genre_scores_gemma":[0.02977494,0.000206044,0.961478,0.0001240128,0.00007489182,0.0005166616,0.00520127,0.0008726419,0.001751508],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01226791,"threshold_uncertainty_score":0.02439302,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2072732320","doi":"10.1007/s10618-005-0019-1","title":"Data Clustering with Partial Supervision","year":2006,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":75,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Cluster analysis; Computer science; Artificial intelligence; Classifier (UML); Pattern recognition (psychology); Fuzzy clustering; Machine learning; Support vector machine; Data mining","authors":[{"name":"Abdelhamid Bouchachia","is_ca":false},{"name":"Witold Pedrycz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07186957610495115,"gpt":0.3322306622859221,"spread":0.260361086180971,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004555407,0.0008085172,0.002184588,0.001374606,0.001302216,0.002090327,0.002616422,0.001203297,0.002848036],"category_scores_gemma":[0.0220838,0.001122146,0.001975679,0.002114159,0.001751276,0.004056206,0.003653304,0.001552107,0.001392676],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008077853,"about_ca_system_score_gemma":0.00264174,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00240993,"about_ca_topic_score_gemma":0.003934134,"domain_scores_codex":[0.9943514,0.002137703,0.0004170933,0.001361777,0.00155833,0.0001736776],"domain_scores_gemma":[0.9855563,0.00413865,0.0007965696,0.007122415,0.002040097,0.0003459355],"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":[0.0009985493,0.0002153772,0.008193932,0.0007827635,0.0006102683,0.0003059329,0.0005634647,0.2593059,0.007974184,0.06632493,0.01252527,0.6421995],"study_design_scores_gemma":[0.00003029683,0.00007733476,0.0007943041,0.00002196082,0.00005532573,0.0001578326,0.00003933338,0.9392262,0.003600049,0.05306476,0.002914346,0.0000183528],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00929492,0.0002416268,0.9879447,0.0002582112,0.00003962943,0.00007452391,0.0002601745,0.001052391,0.0008337929],"genre_scores_gemma":[0.3790835,0.0003808873,0.6139207,0.0001876275,0.0002165162,0.0003732599,0.002162853,0.0002577119,0.003416892],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004555407,"threshold_uncertainty_score":0.02409154,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1999518899","doi":"10.1007/s10618-014-0398-2","title":"Mining outlying aspects on numeric data","year":2015,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":70,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"China Postdoctoral Science Foundation; National Natural Science Foundation of China","keywords":"Subspace topology; Object (grammar); Outlier; Computer science; Data mining; Set (abstract data type); Rank (graph theory); Heuristic; Curse of dimensionality; Measure (data warehouse); Data set; Mathematics; Artificial intelligence; Combinatorics","authors":[{"name":"Lei Duan","is_ca":false},{"name":"Guanting Tang","is_ca":true},{"name":"Jian Pei","is_ca":true},{"name":"James Bailey","is_ca":false},{"name":"Akiko Campbell","is_ca":false},{"name":"Changjie Tang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1626365810308897,"gpt":0.3484467786911583,"spread":0.1858101976602685,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002479302,0.0007806382,0.001220317,0.005686093,0.001044109,0.002504597,0.00141197,0.0009756576,0.0007559061],"category_scores_gemma":[0.02586422,0.0004243968,0.0009766672,0.006963037,0.001159593,0.005039849,0.001985298,0.001819175,0.0003455668],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006256874,"about_ca_system_score_gemma":0.001273179,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001658244,"about_ca_topic_score_gemma":0.002448636,"domain_scores_codex":[0.9962114,0.000426325,0.0005044695,0.0006659655,0.001935187,0.0002566552],"domain_scores_gemma":[0.9726827,0.0121404,0.005294499,0.004979481,0.004311825,0.0005910455],"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":[0.0008550968,0.0003627724,0.4227242,0.0006895618,0.0004446939,0.004577211,0.002760668,0.02138935,0.02432215,0.02675872,0.005293347,0.4898223],"study_design_scores_gemma":[0.00006128171,0.0005341182,0.1254985,0.0004401062,0.0007220343,0.006281441,0.003961196,0.5935636,0.0394789,0.2039662,0.0253551,0.0001376069],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6128554,0.00132491,0.3777225,0.001119199,0.0001301942,0.0001490754,0.002044903,0.001966867,0.002686969],"genre_scores_gemma":[0.8387077,0.0006391647,0.1554639,0.0001726359,0.0001884474,0.00008904442,0.003435644,0.0002063196,0.001097132],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005686093,"threshold_uncertainty_score":0.01311195,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2073184692","doi":"10.1007/s10618-005-1359-6","title":"Data Mining for Inventory Item Selection with Cross-Selling Considerations","year":2005,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":70,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Association rule learning; Computer science; Data mining; Profit (economics); Selection (genetic algorithm); Greedy algorithm; Machine learning; Algorithm; Economics","authors":[{"name":"Raymond Chi-Wing Wong","is_ca":false},{"name":"Ada Wai-Chee Fu","is_ca":false},{"name":"Ke Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1057515397087488,"gpt":0.3499721572955075,"spread":0.2442206175867587,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004446562,0.0009305151,0.001903939,0.004269177,0.001116348,0.003113904,0.001860716,0.0009561267,0.005448673],"category_scores_gemma":[0.01992637,0.0006811997,0.001926769,0.005962026,0.0005774868,0.003994374,0.001203365,0.001300093,0.001060895],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007593615,"about_ca_system_score_gemma":0.001119629,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001160558,"about_ca_topic_score_gemma":0.001762806,"domain_scores_codex":[0.9966975,0.0008399875,0.0004095981,0.000725933,0.001159139,0.0001678758],"domain_scores_gemma":[0.9888958,0.007637458,0.0007471894,0.001247501,0.001276434,0.000195677],"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":[0.001990671,0.001214597,0.1254745,0.001152088,0.001173488,0.002001573,0.0004073056,0.1022444,0.004437984,0.03825939,0.01270839,0.7089356],"study_design_scores_gemma":[0.0001249656,0.000548474,0.02392118,0.0002248961,0.0005565143,0.001923047,0.0004268203,0.8859912,0.006802899,0.0683637,0.0110503,0.00006610526],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2529173,0.004295668,0.7220311,0.001945096,0.0002713809,0.000606527,0.004077391,0.001141816,0.0127137],"genre_scores_gemma":[0.7569119,0.0008123646,0.2336252,0.0002152391,0.0002921006,0.0003090598,0.003915254,0.00007253004,0.003846245],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005448673,"threshold_uncertainty_score":0.023516,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2010971425","doi":"10.1007/s10618-011-0212-3","title":"Measuring the component overlapping in the Gaussian mixture model","year":2011,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":60,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Cluster analysis; Gaussian; Computer science; Algorithm; Set (abstract data type); Component (thermodynamics); Multivariate normal distribution; Mixture model; Measure (data warehouse); Mathematics; Data mining; Multivariate statistics; Pattern recognition (psychology); Artificial intelligence; Machine learning","authors":[{"name":"Haojun Sun","is_ca":false},{"name":"Shengrui Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1637535893682168,"gpt":0.3177989488666944,"spread":0.1540453594984776,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007576704,0.001149678,0.001711906,0.003760576,0.001233904,0.003047348,0.002140683,0.002754769,0.001137866],"category_scores_gemma":[0.04215915,0.0008143449,0.001245828,0.00376109,0.00202863,0.005212673,0.002972289,0.001793928,0.0005117311],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001167714,"about_ca_system_score_gemma":0.001254622,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004889307,"about_ca_topic_score_gemma":0.002713131,"domain_scores_codex":[0.9936507,0.002371164,0.0003064643,0.001168198,0.002032174,0.0004713232],"domain_scores_gemma":[0.9801658,0.01353327,0.001174364,0.002590504,0.002015208,0.0005207778],"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":[0.001772096,0.0003880253,0.05562976,0.0005531026,0.001034749,0.0004554652,0.001660801,0.5435554,0.02667048,0.1018234,0.002965267,0.2634914],"study_design_scores_gemma":[0.00001513588,0.00007522311,0.008059748,0.00002831405,0.0001262048,0.0003239846,0.0001385213,0.9529648,0.004293298,0.0329752,0.0009392739,0.00006025826],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.104692,0.0005797641,0.8925421,0.0001676184,0.00004299302,0.00004140039,0.00009155746,0.0003448396,0.001497724],"genre_scores_gemma":[0.7774403,0.0006471138,0.2202113,0.0001028953,0.00009189534,0.00009057267,0.0004989449,0.0001937566,0.0007231524],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007576704,"threshold_uncertainty_score":0.04006994,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2161479324","doi":"10.1007/s10618-006-0051-9","title":"Privacy-preserving boosting","year":2007,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":60,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Boosting (machine learning); AdaBoost; Computer science; Artificial intelligence; Classifier (UML); Machine learning; Generalization; Bipartite graph; Data mining; Pattern recognition (psychology); Theoretical computer science; Mathematics","authors":[{"name":"Sébastien Gambs","is_ca":true},{"name":"Balázs Kégl","is_ca":true},{"name":"Esma Aı̈meur","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0657895914618925,"gpt":0.3222549855121698,"spread":0.2564653940502773,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004060217,0.0007476427,0.001895287,0.0008971051,0.001255932,0.002063696,0.001708337,0.001366584,0.003865052],"category_scores_gemma":[0.01323786,0.0005508712,0.001217739,0.001304555,0.001503902,0.002571899,0.002429253,0.002514172,0.001421614],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007558589,"about_ca_system_score_gemma":0.001707906,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006353454,"about_ca_topic_score_gemma":0.0006513648,"domain_scores_codex":[0.9963672,0.001438149,0.000144744,0.0006899037,0.0009867486,0.0003730948],"domain_scores_gemma":[0.9929225,0.002628766,0.0002626546,0.00315522,0.0008173888,0.0002133291],"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":[0.000744598,0.0003130811,0.003543685,0.0002645823,0.0002526075,0.0002156442,0.0002486938,0.147548,0.009935724,0.345478,0.02668894,0.4647664],"study_design_scores_gemma":[0.00006688981,0.0001573737,0.0007364447,0.00003164036,0.00008724734,0.0004373708,0.0000347231,0.6199657,0.008300797,0.3552824,0.01487652,0.00002288028],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02431259,0.001078723,0.9609758,0.001239535,0.0002423038,0.0001032765,0.0002841632,0.001143704,0.01061975],"genre_scores_gemma":[0.7255465,0.0009739464,0.2554756,0.0009889362,0.0006337221,0.0001842315,0.0007714563,0.0002834548,0.01514216],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004060217,"threshold_uncertainty_score":0.02147275,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2098162950","doi":"10.1023/b:dami.0000023675.04946.f1","title":"Building Association-Rule Based Sequential Classifiers for Web-Document Prediction","year":2004,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":60,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Association rule learning; Data mining; Web application; Predictive modelling; Web server; Association (psychology); Web mining; Machine learning; Web page; Information retrieval; World Wide Web; The Internet","authors":[{"name":"Qiang Yang","is_ca":true},{"name":"Tianyi Li","is_ca":true},{"name":"Ke Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03959326074880023,"gpt":0.3087525581662589,"spread":0.2691592974174587,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004083299,0.001156716,0.002199924,0.004789432,0.001079285,0.001953871,0.002136444,0.001945255,0.002617349],"category_scores_gemma":[0.01114064,0.0007686554,0.001422182,0.002926238,0.0004499768,0.003062223,0.0009944532,0.002031642,0.00288123],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006365745,"about_ca_system_score_gemma":0.002278228,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006652573,"about_ca_topic_score_gemma":0.008279706,"domain_scores_codex":[0.9978147,0.0003350128,0.0003234788,0.0005845709,0.0007595075,0.0001828173],"domain_scores_gemma":[0.9902829,0.005532743,0.0005306698,0.0007653629,0.002549459,0.0003389814],"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":[0.0007809278,0.0009576234,0.02663937,0.0002982396,0.0003260653,0.0004770278,0.0001398771,0.04796015,0.01179805,0.002455259,0.00864098,0.8995265],"study_design_scores_gemma":[0.00005050786,0.0002488537,0.002274713,0.00003498915,0.0001723718,0.0003086258,0.0000529542,0.978192,0.009758197,0.006144866,0.002733154,0.00002865976],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1102667,0.00119192,0.8774124,0.0004412805,0.0003195434,0.0005861175,0.001494781,0.006149533,0.00213779],"genre_scores_gemma":[0.2958241,0.0006145509,0.6944715,0.0002007328,0.0003006115,0.00052051,0.003836523,0.0001694753,0.004062057],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006652573,"threshold_uncertainty_score":0.02159482,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1964878798","doi":"10.1007/s10618-010-0208-4","title":"Publishing anonymous survey rating data","year":2010,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":56,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Anonymity; Computer science; Database transaction; Graph; Information retrieval; Personally identifiable information; Internet privacy; Data mining; Similarity (geometry); k-anonymity; Computer security; Theoretical computer science; Database; Artificial intelligence","authors":[{"name":"Xiaoxun Sun","is_ca":false},{"name":"Hua Wang","is_ca":false},{"name":"Jiuyong Li","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1226799404888649,"gpt":0.3328434296471469,"spread":0.2101634891582821,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01115351,0.0007161248,0.001801148,0.002947415,0.001202333,0.004463987,0.002085301,0.002076428,0.01468504],"category_scores_gemma":[0.09362465,0.0007991646,0.0009956212,0.006454316,0.001035017,0.004771827,0.002109329,0.001993249,0.01100141],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001570903,"about_ca_system_score_gemma":0.002435655,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00496924,"about_ca_topic_score_gemma":0.006544378,"domain_scores_codex":[0.9800805,0.008225473,0.001305708,0.002386993,0.007009163,0.0009921426],"domain_scores_gemma":[0.9068324,0.02471179,0.005230597,0.0519529,0.01012586,0.001146405],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001674175,0.0004191632,0.03618611,0.0006521273,0.0004533835,0.0006960175,0.000853941,0.04511751,0.003104537,0.2066284,0.250856,0.4533587],"study_design_scores_gemma":[0.0002742399,0.0003153362,0.01492685,0.0002300414,0.0002524203,0.00154869,0.0008024602,0.4050463,0.01153699,0.3285797,0.2362956,0.0001913151],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1277309,0.002428321,0.6864615,0.0146977,0.003380598,0.001016397,0.07800538,0.009019743,0.07725938],"genre_scores_gemma":[0.8230707,0.001378442,0.08293692,0.001140039,0.001396394,0.0006134092,0.03734302,0.0004270402,0.05169406],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9888465,"threshold_uncertainty_score":0.05898613,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2884063665","doi":"10.1007/s10618-018-0578-6","title":"Extreme-value-theoretic estimation of local intrinsic dimensionality","year":2018,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Statistical Mechanics and Entropy","field":"Physics and Astronomy","cited_by":52,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Google (Canada)","funders":"","keywords":"Curse of dimensionality; Discriminative model; Estimator; Outlier; Measure (data warehouse); Intrinsic dimension; Extreme value theory; Mathematics; Random variable; Similarity measure; Similarity (geometry); Dimension (graph theory); Pattern recognition (psychology); Artificial intelligence; Computer science; Data mining; Statistics","authors":[{"name":"Laurent Amsaleg","is_ca":false},{"name":"Oussama Chelly","is_ca":false},{"name":"Teddy Furon","is_ca":false},{"name":"Stéphane Girard","is_ca":false},{"name":"Michael E. Houle","is_ca":false},{"name":"Ken‐ichi Kawarabayashi","is_ca":false},{"name":"Michael Nett","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03719904815149069,"gpt":0.3063922545785057,"spread":0.2691932064270151,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005411464,0.0006044334,0.001389949,0.002131619,0.0005422162,0.002483685,0.001778618,0.001162464,0.000953758],"category_scores_gemma":[0.02922844,0.0004990915,0.0008574643,0.001301216,0.00293691,0.004567659,0.003030165,0.002298801,0.0002262533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001090181,"about_ca_system_score_gemma":0.00112412,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006108496,"about_ca_topic_score_gemma":0.0005215958,"domain_scores_codex":[0.9978435,0.001150244,0.0001098848,0.000340468,0.0004450631,0.000110744],"domain_scores_gemma":[0.9838613,0.01219476,0.001146999,0.001440394,0.0009565104,0.0004000229],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001877125,0.0001534567,0.007792904,0.0003060072,0.0002650622,0.0001449359,0.0002415635,0.4696325,0.003509068,0.4427658,0.002182453,0.07281864],"study_design_scores_gemma":[0.000006814769,0.00002473536,0.0008695321,0.0000190586,0.00001008841,0.00004264629,0.00002493212,0.7657527,0.0006212773,0.2323086,0.0003029045,0.00001666754],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02807209,0.0003606399,0.9701266,0.0004663482,0.00002194704,0.00001272701,0.00008478038,0.0000753368,0.0007795991],"genre_scores_gemma":[0.8638601,0.0009614605,0.1321757,0.0001747823,0.0003460828,0.0001103546,0.000517332,0.00006842877,0.001785683],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005411464,"threshold_uncertainty_score":0.02861887,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2470642288","doi":"10.1007/s10618-016-0475-9","title":"Ensembles of label noise filters: a ranking approach","year":2016,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Computer science; Noise (video); Machine learning; Artificial intelligence; Filter (signal processing); Ranking (information retrieval); Ensemble learning; Process (computing); Data mining; Noise measurement; Pattern recognition (psychology); Noise reduction","authors":[{"name":"Luís P. F. Garcia","is_ca":false},{"name":"Ana Carolina Lorena","is_ca":false},{"name":"Stan Matwin","is_ca":true},{"name":"André C. P. L. F. de Carvalho","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06737345868108362,"gpt":0.3006821942314548,"spread":0.2333087355503712,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01139137,0.002017853,0.004516779,0.005171672,0.001723571,0.004398196,0.003970265,0.00405093,0.004067419],"category_scores_gemma":[0.02845275,0.001031867,0.002121395,0.003534862,0.001113062,0.00463755,0.002207671,0.002848846,0.00193864],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001580606,"about_ca_system_score_gemma":0.002478156,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004438935,"about_ca_topic_score_gemma":0.0078022,"domain_scores_codex":[0.9918705,0.003293922,0.0004680416,0.001200889,0.002518838,0.0006478258],"domain_scores_gemma":[0.9751174,0.013035,0.001196898,0.003335915,0.006530433,0.0007843656],"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":[0.0009338552,0.0008549999,0.007824241,0.00045782,0.001080593,0.000159857,0.0002651995,0.1817589,0.009058356,0.04750815,0.01839218,0.7317058],"study_design_scores_gemma":[0.00005608365,0.0003088283,0.001254962,0.00006021472,0.0002788371,0.0001044751,0.00008531004,0.9445593,0.003973937,0.04571069,0.003547602,0.00005977256],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01138049,0.0008137513,0.9847937,0.0003336211,0.000132451,0.00008978421,0.0001450691,0.0008310222,0.001480059],"genre_scores_gemma":[0.3220713,0.001037976,0.6634712,0.0005594203,0.0008554559,0.000364907,0.001556,0.0006171654,0.009466656],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01139137,"threshold_uncertainty_score":0.06024402,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4376956106","doi":"10.1007/s10618-023-00931-x","title":"On the evaluation of outlier detection and one-class classification: a comparative study of algorithms, model selection, and ensembles","year":2023,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":29,"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; Danmarks Frie Forskningsfond; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Anomaly detection; Outlier; Artificial intelligence; Machine learning; One-class classification; Computer science; Class (philosophy); Selection (genetic algorithm); Pattern recognition (psychology); Support vector machine; Focus (optics); Data mining","authors":[{"name":"Henrique O. Marques","is_ca":false},{"name":"Lorne Swersky","is_ca":true},{"name":"Jörg Sander","is_ca":true},{"name":"Ricardo J. G. B. Campello","is_ca":false},{"name":"Arthur Zimek","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2707064289841571,"gpt":0.3888650291681965,"spread":0.1181586001840393,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04101722,0.003863413,0.004573241,0.008709935,0.001422035,0.003366004,0.003098741,0.003831777,0.0008239626],"category_scores_gemma":[0.09605254,0.0006378058,0.002631884,0.006286619,0.001879621,0.005892192,0.003514807,0.003190109,0.0005435743],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002187158,"about_ca_system_score_gemma":0.001853941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004784373,"about_ca_topic_score_gemma":0.003499206,"domain_scores_codex":[0.9618884,0.01901599,0.002857226,0.004973297,0.01045484,0.0008102549],"domain_scores_gemma":[0.8621141,0.1023204,0.007322048,0.01341364,0.01349314,0.001336597],"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":[0.0021565,0.001057644,0.03930902,0.001744613,0.003361309,0.0001804658,0.0007304281,0.4652346,0.003965843,0.008263889,0.008589461,0.4654063],"study_design_scores_gemma":[0.0001118703,0.001995687,0.01589335,0.0002948157,0.000465308,0.0004255994,0.0004249856,0.9605491,0.007019425,0.009382459,0.003268826,0.000168507],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3847287,0.0416504,0.5573995,0.001881202,0.00108528,0.0007974702,0.001549934,0.003612157,0.007295324],"genre_scores_gemma":[0.7526343,0.003913577,0.2360895,0.000557973,0.0005449363,0.0004566039,0.004129233,0.0006236225,0.001050232],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.04101722,"threshold_uncertainty_score":0.2169224,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2969639627","doi":"10.1007/s10618-019-00651-1","title":"A unified view of density-based methods forsemi-supervised clustering and classification","year":2019,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","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 Alberta","funders":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Natural Sciences and Engineering Research Council of Canada; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior","keywords":"Cluster analysis; Computer science; Artificial intelligence; Semi-supervised learning; Supervised learning; Graph; Machine learning; Correlation clustering; Data stream clustering; Data mining; Pattern recognition (psychology); CURE data clustering algorithm; Artificial neural network; Theoretical computer science","authors":[{"name":"Jadson Castro Gertrudes","is_ca":false},{"name":"Arthur Zimek","is_ca":false},{"name":"Jörg Sander","is_ca":true},{"name":"Ricardo J. G. B. Campello","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09721739508829468,"gpt":0.386530902237729,"spread":0.2893135071494343,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008548399,0.001647976,0.002592206,0.005399259,0.001763481,0.004975011,0.004579043,0.002704513,0.001668639],"category_scores_gemma":[0.01801204,0.001197969,0.002155576,0.004439699,0.004903835,0.006716245,0.004768807,0.005066001,0.001331289],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002352287,"about_ca_system_score_gemma":0.002264068,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002798658,"about_ca_topic_score_gemma":0.002541274,"domain_scores_codex":[0.9917488,0.003681258,0.000419067,0.001305617,0.002639085,0.0002061383],"domain_scores_gemma":[0.988817,0.005699157,0.0006955759,0.00192801,0.002537811,0.0003224912],"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":[0.00006104752,0.0001008203,0.001305696,0.0005383142,0.0002341418,0.0001408805,0.0005996998,0.1447231,0.002896853,0.663209,0.005714659,0.1804758],"study_design_scores_gemma":[0.00001353848,0.00005474525,0.000442032,0.00008716229,0.00004147612,0.000202532,0.0001133539,0.6448022,0.001731306,0.3420829,0.01036331,0.0000654121],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.000402944,0.0003478069,0.9983767,0.000215424,0.00002926632,0.00002473975,0.00002218305,0.00007569274,0.0005052448],"genre_scores_gemma":[0.0800152,0.002290327,0.913923,0.0004141931,0.0006461876,0.0004792611,0.0003459374,0.0001675972,0.001718285],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008548399,"threshold_uncertainty_score":0.04520875,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2545851563","doi":"10.1007/s10618-016-0481-y","title":"Multiple Bayesian discriminant functions for high-dimensional massive data classification","year":2016,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"Université Pierre et Marie Curie; Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Artificial intelligence; Pattern recognition (psychology); Weighting; Naive Bayes classifier; Machine learning; Computer science; Linear discriminant analysis; Discriminant; Bayesian probability; Classifier (UML); Bayesian inference; Feature (linguistics); Clustering high-dimensional data; Data mining; Mathematics; Support vector machine; Cluster analysis","authors":[{"name":"Jianfei Zhang","is_ca":true},{"name":"Shengrui Wang","is_ca":true},{"name":"Lifei Chen","is_ca":false},{"name":"Patrick Gallinari","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09958120661649893,"gpt":0.3066862726524197,"spread":0.2071050660359207,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004874636,0.001185734,0.002947334,0.002837059,0.001277019,0.001906888,0.002753406,0.001775131,0.002012838],"category_scores_gemma":[0.01774734,0.0008673188,0.001478984,0.003328402,0.00107623,0.002652093,0.003052948,0.003304627,0.0016813],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001033959,"about_ca_system_score_gemma":0.001680133,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003732892,"about_ca_topic_score_gemma":0.004748477,"domain_scores_codex":[0.9968488,0.001295213,0.0002507875,0.0004791727,0.0009120027,0.0002141128],"domain_scores_gemma":[0.9925265,0.004754771,0.0004496776,0.001180667,0.0008678994,0.000220452],"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":[0.000317114,0.0003028573,0.003776158,0.0002740857,0.000254078,0.000130848,0.0001578748,0.2132872,0.003169731,0.04055574,0.01131844,0.7264558],"study_design_scores_gemma":[0.00001362006,0.00001666878,0.0004817711,0.00001693885,0.00001561716,0.00003520438,0.0000239423,0.9545681,0.0005560768,0.04314188,0.001114866,0.00001538376],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006363886,0.000892458,0.9911284,0.000395503,0.0000520763,0.00003546442,0.0002135057,0.0006317725,0.0002868868],"genre_scores_gemma":[0.223475,0.001134077,0.7701854,0.0002778346,0.0003417433,0.0004227355,0.001455126,0.0001782909,0.002529786],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004874636,"threshold_uncertainty_score":0.02577984,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2158459641","doi":"10.1007/s10618-015-0408-z","title":"Tractome: a visual data mining tool for brain connectivity analysis","year":2015,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Neuroimaging Techniques and Applications","field":"Medicine","cited_by":24,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Computer science; Visualization; Scalability; Artificial intelligence; Visual analytics; Process (computing); Set (abstract data type); Data visualization; Data set; Data mining; Machine learning; Human–computer interaction; Database","authors":[{"name":"Diana Porro‐Muñoz","is_ca":false},{"name":"Emanuele Olivetti","is_ca":false},{"name":"Nusrat Sharmin","is_ca":false},{"name":"Bao T. Nguyen","is_ca":false},{"name":"Eleftherios Garyfallidis","is_ca":true},{"name":"Paolo Avesani","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2789212773211442,"gpt":0.467167125821036,"spread":0.1882458484998917,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001038947,0.001656798,0.0009319311,0.005017964,0.0006008502,0.001911336,0.001470279,0.0009069865,0.02602215],"category_scores_gemma":[0.007130855,0.0005701481,0.001577642,0.003161358,0.0003996253,0.00150364,0.00220321,0.001213078,0.004632767],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005177675,"about_ca_system_score_gemma":0.001315247,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004312967,"about_ca_topic_score_gemma":0.007250422,"domain_scores_codex":[0.9996536,0.00005980137,0.00005119724,0.00009780454,0.000107702,0.00002986613],"domain_scores_gemma":[0.998175,0.001099644,0.0002135406,0.0002120114,0.0001947953,0.0001049829],"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":[0.001439671,0.0002349923,0.007688174,0.004255372,0.00105798,0.001519873,0.0012351,0.02010508,0.03142702,0.03783346,0.3085219,0.5846815],"study_design_scores_gemma":[0.0004486093,0.0004073707,0.01329636,0.0008968294,0.0006578374,0.003441594,0.0005501052,0.4320268,0.05326079,0.1866038,0.308052,0.000357892],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.01031791,0.0008727956,0.7605737,0.0006893969,0.0002120417,0.0004105578,0.05882861,0.1636234,0.004471512],"genre_scores_gemma":[0.1107107,0.001371392,0.8281956,0.0004306613,0.0001438622,0.002161629,0.03960565,0.01171208,0.005668394],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.02602215,"threshold_uncertainty_score":0.08705282,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2005934890","doi":"10.1007/s10618-013-0336-8","title":"Clustering categorical data in projected spaces","year":2013,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Cluster analysis; Categorical variable; Computer science; Data mining; Outlier; CURE data clustering algorithm; Curse of dimensionality; Clustering high-dimensional data; Correlation clustering; Canopy clustering algorithm; Data stream clustering; Anomaly detection; Probabilistic logic; Set (abstract data type); Artificial intelligence; Pattern recognition (psychology); Machine learning","authors":[{"name":"Mohamed Bouguessa","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09257415986957443,"gpt":0.3536923389834265,"spread":0.261118179113852,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003565447,0.0006601155,0.00144292,0.003296275,0.0009856101,0.004170629,0.001682823,0.0008642421,0.00182067],"category_scores_gemma":[0.01966585,0.0004967592,0.001414593,0.005951377,0.001869137,0.003861298,0.00321019,0.002012498,0.0006795465],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001066883,"about_ca_system_score_gemma":0.00124046,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002782043,"about_ca_topic_score_gemma":0.002104797,"domain_scores_codex":[0.9939646,0.002668276,0.000378107,0.0008925467,0.001863327,0.0002331087],"domain_scores_gemma":[0.9905229,0.00468479,0.00065285,0.001505294,0.00226564,0.0003685236],"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":[0.0006760267,0.0003444984,0.01595098,0.0007624481,0.0004116018,0.000255947,0.001876909,0.2193246,0.006321073,0.2825914,0.005340605,0.4661439],"study_design_scores_gemma":[0.00002360792,0.0001145821,0.002372232,0.00006407911,0.00003795208,0.0001662636,0.0005937868,0.7229326,0.001872669,0.268724,0.003053106,0.00004528843],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02668364,0.0002873314,0.9716309,0.0002125753,0.00003736151,0.00005623259,0.0002977176,0.0002017069,0.0005925907],"genre_scores_gemma":[0.4011313,0.0008362117,0.5930611,0.0001097908,0.0001611146,0.0003080411,0.002352373,0.0001089263,0.001931086],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004170629,"threshold_uncertainty_score":0.01885611,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2154217032","doi":"10.1007/s10618-009-0133-6","title":"Fast incremental mining of web sequential patterns with PLWAP tree","year":2009,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"","keywords":"Computer science; Data mining; Tree (set theory); Metadata; Web mining; Database; Web log analysis software; Web page; Information retrieval; Web navigation; World Wide Web; Static web page","authors":[{"name":"C. I. Ezeife","is_ca":true},{"name":"Yi Liu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0352025202348579,"gpt":0.2851548256075449,"spread":0.249952305372687,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008474107,0.0005587126,0.0009631357,0.003480332,0.0007403493,0.001195083,0.001535092,0.0005985381,0.00185684],"category_scores_gemma":[0.005530823,0.0005411349,0.0009366437,0.003947287,0.0002615497,0.002246692,0.001079112,0.0008036838,0.001156178],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002513818,"about_ca_system_score_gemma":0.001177019,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003152669,"about_ca_topic_score_gemma":0.005857447,"domain_scores_codex":[0.9991241,0.0001100463,0.00009772813,0.0001655412,0.0004113292,0.00009108273],"domain_scores_gemma":[0.997486,0.001091949,0.000206931,0.0004075131,0.0006813127,0.0001262469],"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":[0.0006190049,0.0002977216,0.01444255,0.0005609684,0.0001986943,0.0008946408,0.0002249265,0.02082615,0.01736717,0.004538511,0.01429261,0.9257371],"study_design_scores_gemma":[0.0001160832,0.0002801112,0.005052134,0.00005567854,0.0001873411,0.001439285,0.0001957757,0.9434766,0.015534,0.0214714,0.01215787,0.00003369222],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1037882,0.001170718,0.8797337,0.000327609,0.000162385,0.0004484724,0.003385322,0.008791653,0.002192128],"genre_scores_gemma":[0.2790891,0.0004487055,0.7112548,0.00009837242,0.00007398587,0.0003044787,0.006298919,0.0002422971,0.002189409],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003480332,"threshold_uncertainty_score":0.00626862,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2002041176","doi":"10.1007/s10618-006-0046-6","title":"Relational peculiarity-oriented mining","year":2007,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Rough Sets and Fuzzy Logic","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Regina","funders":"National Natural Science Foundation of China","keywords":"Relevance (law); Computer science; Identification (biology); Relational database; Data mining; Task (project management); Focus (optics); Information retrieval; Engineering","authors":[{"name":"Muneaki Ohshima","is_ca":false},{"name":"Ning Zhong","is_ca":false},{"name":"Yiyu Yao","is_ca":true},{"name":"Chunnian Liu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0726774133759716,"gpt":0.2994970904899545,"spread":0.2268196771139829,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001860904,0.0004719916,0.000691908,0.003276577,0.0008838186,0.002382839,0.00143956,0.0004301407,0.002653915],"category_scores_gemma":[0.006960245,0.000258581,0.0009505679,0.004525611,0.00128822,0.003861981,0.001682834,0.0008490568,0.0006761827],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005451264,"about_ca_system_score_gemma":0.0009504379,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008292221,"about_ca_topic_score_gemma":0.00117569,"domain_scores_codex":[0.9986235,0.0002614072,0.000126657,0.0004161944,0.0004743974,0.00009786042],"domain_scores_gemma":[0.9968753,0.00102645,0.0004365814,0.0009697368,0.0005567204,0.0001351499],"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":[0.0002776365,0.0002014625,0.07578823,0.0009182084,0.0004752068,0.001355623,0.00191177,0.00951111,0.01097632,0.4433631,0.00659431,0.448627],"study_design_scores_gemma":[0.0000492406,0.0001533085,0.02735781,0.0001900216,0.0006485448,0.003217984,0.001774687,0.1090899,0.0118009,0.7947787,0.05085241,0.00008637761],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1945401,0.002184992,0.7629669,0.001816306,0.0001630778,0.0003906574,0.001680915,0.001163384,0.03509361],"genre_scores_gemma":[0.7460886,0.001788391,0.2442667,0.0003510022,0.0001900172,0.0001274348,0.002159073,0.0001009827,0.0049279],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003276577,"threshold_uncertainty_score":0.009841502,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2894980798","doi":"10.1007/s10618-018-0592-8","title":"Classification with label noise: a Markov chain sampling framework","year":2018,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Australian Research Council","keywords":"Markov chain; Computer science; Probabilistic logic; Correctness; Artificial intelligence; Noise (video); Markov chain Monte Carlo; Pattern recognition (psychology); Sampling (signal processing); Machine learning; Probability distribution; Variable-order Markov model; Hidden Markov model; Algorithm; Markov model; Bayesian probability; Mathematics; Statistics","authors":[{"name":"Zijin Zhao","is_ca":true},{"name":"Lingyang Chu","is_ca":true},{"name":"Dacheng Tao","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0757148656264769,"gpt":0.3395744944047306,"spread":0.2638596287782538,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02896227,0.001439573,0.005031454,0.003985937,0.002466375,0.00513718,0.00784435,0.00541053,0.005020238],"category_scores_gemma":[0.06281885,0.002310308,0.00358787,0.003514089,0.005642083,0.007998941,0.004813591,0.005509546,0.001145473],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003415514,"about_ca_system_score_gemma":0.003248833,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00917977,"about_ca_topic_score_gemma":0.008347616,"domain_scores_codex":[0.988277,0.006926228,0.0004267306,0.001909179,0.001814731,0.000646107],"domain_scores_gemma":[0.9125906,0.07676171,0.002204055,0.004733904,0.002601099,0.001108474],"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":[0.0006115454,0.0004234712,0.005263226,0.0002752563,0.0003784273,0.0003498978,0.0006337821,0.5302946,0.0009468098,0.3883322,0.004813729,0.06767718],"study_design_scores_gemma":[0.00003459049,0.00001952617,0.0001118394,0.00001706898,0.00002551796,0.00002311859,0.00001186187,0.9291798,0.0001343774,0.07015195,0.0002780841,0.00001220669],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01073152,0.0002918735,0.9872181,0.0006165838,0.00006886005,0.0001637442,0.0001299909,0.000236208,0.0005431006],"genre_scores_gemma":[0.5377534,0.001184787,0.4467075,0.0009876913,0.001057238,0.001650913,0.001933767,0.0004114807,0.008313198],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02896227,"threshold_uncertainty_score":0.153169,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2950515936","doi":"10.1007/s10618-017-0497-y","title":"Adjusting for scorekeeper bias in NBA box scores","year":2017,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":12,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Statistics; Popularity; Basketball; Computer science; Generosity; Econometrics; Psychology; Mathematics; Geography; Social psychology","authors":[{"name":"Matthew van Bommel","is_ca":true},{"name":"Luke Bornn","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2488841393248247,"gpt":0.3329720170911042,"spread":0.08408787776627952,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.06616323,0.001056761,0.001949132,0.002010254,0.001739323,0.003369816,0.003788523,0.001920362,0.008872538],"category_scores_gemma":[0.2382367,0.0005384845,0.002353433,0.003898535,0.001530655,0.002870517,0.001881154,0.003215345,0.00311347],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001436487,"about_ca_system_score_gemma":0.003919216,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03723755,"about_ca_topic_score_gemma":0.040982,"domain_scores_codex":[0.9587287,0.0230891,0.003160605,0.007471094,0.005369918,0.002180672],"domain_scores_gemma":[0.819275,0.1155889,0.012511,0.03347846,0.01620876,0.002937843],"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":[0.001597598,0.0003127551,0.7374026,0.0003764956,0.006069344,0.0004728929,0.0009089806,0.02529887,0.002544199,0.024441,0.04988135,0.150694],"study_design_scores_gemma":[0.0007366116,0.000824332,0.515743,0.0004969796,0.005303354,0.0007429105,0.001603539,0.2851343,0.01537845,0.09977704,0.07401,0.0002496069],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4678356,0.002705193,0.4843717,0.005873238,0.003045118,0.0005180905,0.0110476,0.004615198,0.01998832],"genre_scores_gemma":[0.9094836,0.0001700507,0.07022471,0.001204394,0.0004140549,0.0002742123,0.004269942,0.0006257336,0.01333336],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.06616323,"threshold_uncertainty_score":0.3499088,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2159694214","doi":"10.1007/s10618-005-0013-7","title":"Discovering Classification from Data of Multiple Sources","year":2006,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"Hong Kong University of Science and Technology","keywords":"Computer science; Class (philosophy); Artificial intelligence; Perspective (graphical); Machine learning; Supervised learning; Test data; Labeled data; Semi-supervised learning; Data mining; Artificial neural network","authors":[{"name":"Charles X. Ling","is_ca":true},{"name":"Qiang Yang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09401072320475932,"gpt":0.3028926707297767,"spread":0.2088819475250174,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005624347,0.001409289,0.002243691,0.01830985,0.001482627,0.006336865,0.002386163,0.002665503,0.001637237],"category_scores_gemma":[0.02580046,0.0008612662,0.002122977,0.01720483,0.001049047,0.008512229,0.002439776,0.002515985,0.002108492],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001028792,"about_ca_system_score_gemma":0.001895905,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002427358,"about_ca_topic_score_gemma":0.003026251,"domain_scores_codex":[0.9932795,0.001063304,0.0007128327,0.001352261,0.003217558,0.0003745958],"domain_scores_gemma":[0.9758523,0.01339128,0.002098589,0.003943132,0.00429087,0.0004238674],"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":[0.001166684,0.001317031,0.1555344,0.001944462,0.00129095,0.002652186,0.001920364,0.007674853,0.02687825,0.006253668,0.0127994,0.7805677],"study_design_scores_gemma":[0.000452867,0.00126247,0.1515622,0.002250052,0.007240615,0.0065847,0.01008784,0.4500056,0.1169482,0.1714756,0.08161869,0.0005113019],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3314754,0.006971492,0.6293526,0.004597206,0.000682692,0.0009879175,0.01468221,0.004189003,0.007061521],"genre_scores_gemma":[0.492104,0.003294614,0.4794604,0.0004703014,0.0007005545,0.0008244904,0.02004827,0.0003483991,0.002748897],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01830985,"threshold_uncertainty_score":0.0297448,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2032708445","doi":"10.1007/s10618-007-0086-6","title":"Classification trees and decision-analytic feedforward control: a case study from the video game industry","year":2008,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Modeling, Simulation, and Optimization","field":"Mathematics","cited_by":11,"is_retracted":false,"has_abstract":false,"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","keywords":"Computer science; Feed forward; Video game; Model predictive control; Decision tree; Probabilistic logic; Control (management); Machine learning; Artificial intelligence; Data mining; Engineering; Multimedia; Control engineering","authors":[{"name":"Michael Brydon","is_ca":true},{"name":"Andrew Gemino","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1710043822641238,"gpt":0.3681496386951982,"spread":0.1971452564310744,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003369571,0.0005214837,0.0005452249,0.001063075,0.001148812,0.00139048,0.001176165,0.002236139,0.002225311],"category_scores_gemma":[0.01367834,0.0002066588,0.0004771051,0.001773018,0.001007734,0.001307702,0.000605438,0.00121194,0.0002009264],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001446452,"about_ca_system_score_gemma":0.001374395,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02095323,"about_ca_topic_score_gemma":0.02221908,"domain_scores_codex":[0.998607,0.000781849,0.00005300577,0.0001079097,0.0003177418,0.0001325948],"domain_scores_gemma":[0.984251,0.01389358,0.0004150398,0.0003222759,0.0008437431,0.0002744391],"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":[0.001863624,0.002239802,0.0348111,0.0004914645,0.0001529014,0.003497411,0.002154534,0.686238,0.002482614,0.06207521,0.005813992,0.1981793],"study_design_scores_gemma":[0.0001222872,0.0003949334,0.005612947,0.00003814531,0.00003646118,0.0002242307,0.001032475,0.9609588,0.001357015,0.02733493,0.002857411,0.00003033417],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8694198,0.0007729685,0.118325,0.001733012,0.0000496275,0.0003061053,0.0002145981,0.0001453817,0.009033514],"genre_scores_gemma":[0.9719302,0.0001909149,0.02574265,0.00006277556,0.00001412459,0.000063299,0.00009151342,0.00001475199,0.00188986],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02095323,"threshold_uncertainty_score":0.04166257,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3099831365","doi":"10.1007/s10618-015-0426-x","title":"Generalization of clustering agreements and distances for overlapping clusters and network communities","year":2015,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Cluster analysis; Generalization; Measure (data warehouse); Context (archaeology); Single-linkage clustering; Correlation clustering; Rand index; Contingency table; Cluster (spacecraft)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.1193435223284324,"gpt":0.3518487569469944,"spread":0.232505234618562,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0154444,0.001322212,0.002192988,0.007750622,0.003083277,0.004336335,0.005461771,0.002765425,0.003525379],"category_scores_gemma":[0.1107473,0.001108027,0.002073954,0.006505052,0.004634399,0.01268965,0.005332076,0.004982627,0.0008704209],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00322626,"about_ca_system_score_gemma":0.001788684,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003681614,"about_ca_topic_score_gemma":0.002954978,"domain_scores_codex":[0.9907832,0.002970137,0.0006410959,0.002855322,0.002322295,0.0004279064],"domain_scores_gemma":[0.9243514,0.04660686,0.007830225,0.01020455,0.008800325,0.002206717],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002208557,0.0001777235,0.01006655,0.0003943261,0.0002112763,0.0002265674,0.002220211,0.1121589,0.001464591,0.7679328,0.00482158,0.1001046],"study_design_scores_gemma":[0.00002094977,0.00004606771,0.002209862,0.00006790943,0.00005217673,0.0002462538,0.0004243967,0.2919247,0.0005645145,0.7020565,0.002345444,0.00004130535],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06791712,0.0008079784,0.9245141,0.0007199816,0.00008862452,0.0001493441,0.0005209373,0.0003231484,0.004958825],"genre_scores_gemma":[0.6191849,0.001427687,0.3716182,0.0003080779,0.0003905653,0.000517234,0.001755068,0.0003088533,0.004489479],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0154444,"threshold_uncertainty_score":0.08167875,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2755646241","doi":"10.1007/s10618-017-0540-z","title":"Bounding the difference between RankRC and RankSVM and application to multi-level rare class kernel ranking","year":2017,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Machine Learning and ELM","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Kernel method; Kernel (algebra); Computer science; Bounding overwatch; Mathematics; Projection (relational algebra); Algorithm; Support vector machine; Machine learning; Artificial intelligence; Combinatorics","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":0.08877169197371115,"gpt":0.3433036432579842,"spread":0.254531951284273,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008853575,0.001260183,0.002535643,0.001986808,0.0009590298,0.004067407,0.003753462,0.003062624,0.0045445],"category_scores_gemma":[0.05491377,0.0005787666,0.0007928012,0.001561801,0.001896684,0.004606394,0.004414476,0.003896018,0.001473225],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0023827,"about_ca_system_score_gemma":0.003570604,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004332137,"about_ca_topic_score_gemma":0.007375269,"domain_scores_codex":[0.9911864,0.003322938,0.0003973824,0.001115009,0.00310127,0.0008769611],"domain_scores_gemma":[0.9696913,0.02090064,0.001283831,0.003357888,0.003907166,0.0008590836],"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":[0.00125598,0.0004879934,0.004441522,0.0005013533,0.0001681056,0.0002932854,0.0002708901,0.4009765,0.0116458,0.1600771,0.01950176,0.4003796],"study_design_scores_gemma":[0.00001269763,0.00003502067,0.0002463806,0.00001156865,0.000007531867,0.00005018274,0.00001427613,0.9816127,0.001293369,0.01596061,0.000743632,0.0000120224],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0341565,0.002418104,0.955542,0.001301341,0.0003285611,0.00007412585,0.000175958,0.002872879,0.003130517],"genre_scores_gemma":[0.5677587,0.0005797714,0.4228218,0.0004097354,0.0004014258,0.0001794815,0.0005218902,0.0008164531,0.006510762],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008853575,"threshold_uncertainty_score":0.04682273,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4283736551","doi":"10.1007/s10618-022-00843-2","title":"VPint: value propagation-based spatial interpolation","year":2022,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Image Processing Techniques","field":"Computer Science","cited_by":7,"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":"Horizon 2020; Nederlandse Organisatie voor Wetenschappelijk Onderzoek","keywords":"Interpolation (computer graphics); Multivariate interpolation; Value (mathematics); Computer science; Mathematics; Algorithm; Bilinear interpolation; Artificial intelligence; Statistics","authors":[{"name":"Laurens Arp","is_ca":false},{"name":"Mitra Baratchi","is_ca":false},{"name":"Holger H. Hoos","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03145311805954926,"gpt":0.3022289843732606,"spread":0.2707758663137113,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003794094,0.001089404,0.001531529,0.001836508,0.0006742831,0.00132737,0.004094868,0.001107867,0.00204281],"category_scores_gemma":[0.01402578,0.0007077931,0.001500084,0.003037854,0.0009199248,0.002967392,0.003344552,0.002915648,0.0007192437],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001251624,"about_ca_system_score_gemma":0.00276736,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.013225,"about_ca_topic_score_gemma":0.01310163,"domain_scores_codex":[0.998276,0.0004671154,0.0001275674,0.0003824217,0.0005903445,0.0001565614],"domain_scores_gemma":[0.9946201,0.002736199,0.0004525731,0.0009780877,0.001006938,0.0002061508],"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":[0.0003721181,0.0002338002,0.005047313,0.0002367676,0.0001685562,0.0001805132,0.000311564,0.5564292,0.003299096,0.01819194,0.006415864,0.4091133],"study_design_scores_gemma":[0.00001789934,0.00003770056,0.000196513,0.00001050847,0.00001195505,0.00003275897,0.00002026416,0.9882821,0.00118719,0.008984528,0.001207319,0.00001127078],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009070682,0.0002491823,0.9871058,0.000160679,0.00006606719,0.00005193077,0.0002147086,0.002500392,0.0005804752],"genre_scores_gemma":[0.2844864,0.0003859142,0.7105706,0.0002313998,0.00009974856,0.0001977409,0.001410687,0.0003974989,0.002219925],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.013225,"threshold_uncertainty_score":0.02629602,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3199460270","doi":"10.1007/s10618-021-00794-0","title":"A Lagrangian-based score for assessing the quality of pairwise constraints in semi-supervised clustering","year":2021,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Polytechnique Montréal; Group for Research in Decision Analysis","funders":"","keywords":"Cluster analysis; Pairwise comparison; Constrained clustering; Constraint (computer-aided design); Computer science; Data mining; Machine learning; Key (lock); Artificial intelligence; Correlation clustering; Mathematics; CURE data clustering algorithm","authors":[{"name":"Rodrigo Randel","is_ca":true},{"name":"Daniel Aloise","is_ca":true},{"name":"Simon J. Blanchard","is_ca":false},{"name":"Alain Hertz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1432772631699324,"gpt":0.4071249385868225,"spread":0.2638476754168901,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01602347,0.001441052,0.002742829,0.006182106,0.001795741,0.003748446,0.003964273,0.003704021,0.002633402],"category_scores_gemma":[0.05293308,0.0006531999,0.001552213,0.004523213,0.003716595,0.004442091,0.003937851,0.002532433,0.0007130142],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002280107,"about_ca_system_score_gemma":0.003606672,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003588218,"about_ca_topic_score_gemma":0.004316756,"domain_scores_codex":[0.986406,0.005164179,0.00112966,0.001355385,0.005344044,0.0006007172],"domain_scores_gemma":[0.9651881,0.01971111,0.003819011,0.002310764,0.007411675,0.001559316],"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":[0.001416416,0.0006294053,0.03132525,0.00108296,0.001120738,0.00045871,0.0005473424,0.6068634,0.01169892,0.06800734,0.01257355,0.2642759],"study_design_scores_gemma":[0.0000809103,0.0002673114,0.002241407,0.00005343985,0.00006120275,0.0001639113,0.00009352922,0.9710988,0.001951778,0.02332013,0.0006028145,0.00006478135],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04499508,0.0003877762,0.9511561,0.0004508108,0.00008828169,0.0001691873,0.0004561699,0.0005238377,0.001772753],"genre_scores_gemma":[0.5308143,0.0002264346,0.4647467,0.0002531692,0.0001788002,0.000344751,0.00184731,0.0002314432,0.001356971],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01602347,"threshold_uncertainty_score":0.08474123,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2061678866","doi":"10.1007/s10618-013-0330-1","title":"What distinguish one from its peers in social networks?","year":2013,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":5,"is_retracted":false,"has_abstract":false,"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 Taiwan University; National Science Council; Intel Corporation","keywords":"Uniqueness; Identification (biology); Computer science; Task (project management); Vertex (graph theory); Group (periodic table); Social network (sociolinguistics); Social network analysis; Theoretical computer science; Data mining; Machine learning; Graph; Mathematics; World Wide Web; Social media; Engineering","authors":[{"name":"Yi‐Chen Lo","is_ca":false},{"name":"Jhao-Yin Li","is_ca":false},{"name":"Mi-Yen Yeh","is_ca":false},{"name":"Shou-De Lin","is_ca":false},{"name":"Jian Pei","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05195597983772818,"gpt":0.3072964131251066,"spread":0.2553404332873784,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001468263,0.0002858412,0.0007099339,0.003003258,0.00117197,0.002751229,0.0008994996,0.001306755,0.001658982],"category_scores_gemma":[0.01886651,0.0001836376,0.0002759387,0.001956194,0.00116112,0.007627993,0.001127994,0.000723456,0.0005674022],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005856529,"about_ca_system_score_gemma":0.0004591098,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003341157,"about_ca_topic_score_gemma":0.00512922,"domain_scores_codex":[0.998632,0.0003724634,0.000084199,0.0003655734,0.0003949452,0.0001507315],"domain_scores_gemma":[0.9900705,0.006118118,0.001436589,0.0004882696,0.001032776,0.0008538127],"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":[0.0006488349,0.0004314742,0.5245116,0.0009541348,0.0005968033,0.001257381,0.008587329,0.005592626,0.00645833,0.08334793,0.01646212,0.3511514],"study_design_scores_gemma":[0.0001481188,0.0004397196,0.2996818,0.0006721703,0.001208348,0.003845627,0.0498738,0.1246063,0.009669505,0.453765,0.05585112,0.0002384439],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9245435,0.00244963,0.04193819,0.007628592,0.0002132625,0.0001401831,0.00145358,0.0001946837,0.02143837],"genre_scores_gemma":[0.991396,0.0005264578,0.006396481,0.0001797848,0.0001527818,0.00003700674,0.0003648437,0.00002186907,0.0009247494],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003341157,"threshold_uncertainty_score":0.007764995,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2106700114","doi":"10.1007/s10618-012-0257-y","title":"Guest editorial: special issue on a decade of mining the Web","year":2012,"lang":"en","type":"editorial","venue":"Data Mining and Knowledge Discovery","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Web mining; World Wide Web; Computer science; Data science; Domain (mathematical analysis); Web intelligence; Web development; The Internet; Web page","authors":[{"name":"Myra Spiliopoulou","is_ca":false},{"name":"Bamshad Mobasher","is_ca":false},{"name":"Olfa Nasraoui","is_ca":false},{"name":"Osmar R. Zai͏̈ane","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02640097352747741,"gpt":0.2997039996304098,"spread":0.2733030261029324,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007987171,0.003699088,0.004478125,0.006782597,0.003068111,0.01090008,0.003323148,0.01065086,0.01760282],"category_scores_gemma":[0.03096463,0.001312182,0.002709245,0.00275888,0.002650972,0.004752626,0.001847062,0.01611575,0.01218007],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002525985,"about_ca_system_score_gemma":0.003091686,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00112043,"about_ca_topic_score_gemma":0.004033853,"domain_scores_codex":[0.993512,0.001036706,0.0007581705,0.0008912482,0.003322983,0.0004788482],"domain_scores_gemma":[0.9627004,0.01195161,0.002811556,0.0009171404,0.0161129,0.005506365],"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":[0.00003131239,0.00001134243,0.00002716299,0.0001651208,0.00001873718,0.00007309381,0.0000070745,0.00001010468,0.00004584702,0.0001387552,0.9959834,0.003487969],"study_design_scores_gemma":[0.00005702593,0.00003633653,0.0003634446,0.000456604,0.00008241004,0.0003180038,0.00004905686,0.0001532839,0.0001744712,0.0009164255,0.997372,0.00002079497],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.00002748631,0.005134751,0.0001220818,0.02462726,0.969138,0.00001169469,0.00004997384,0.0000426066,0.0008461766],"genre_scores_gemma":[0.0002774598,0.003641319,0.0001174415,0.009266715,0.9809971,0.00001451077,0.00003711291,0.00003867278,0.005609628],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.01760282,"threshold_uncertainty_score":0.05888736,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4239981520","doi":"10.1007/s10618-009-0144-3","title":"Fast incremental mining of web sequential patterns with PLWAP tree","year":2009,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"","keywords":"Computer science; Data mining; Tree (set theory); Web mining; Information retrieval; World Wide Web; Mathematics; Web service","authors":[{"name":"C. I. Ezeife","is_ca":true},{"name":"Yi Liu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0352025202348579,"gpt":0.2851548256075449,"spread":0.249952305372687,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008474107,0.0005587126,0.0009631357,0.003480332,0.0007403493,0.001195083,0.001535092,0.0005985381,0.00185684],"category_scores_gemma":[0.005530823,0.0005411349,0.0009366437,0.003947287,0.0002615497,0.002246692,0.001079112,0.0008036838,0.001156178],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002513818,"about_ca_system_score_gemma":0.001177019,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003152669,"about_ca_topic_score_gemma":0.005857447,"domain_scores_codex":[0.9991241,0.0001100463,0.00009772813,0.0001655412,0.0004113292,0.00009108273],"domain_scores_gemma":[0.997486,0.001091949,0.000206931,0.0004075131,0.0006813127,0.0001262469],"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":[0.0006190049,0.0002977216,0.01444255,0.0005609684,0.0001986943,0.0008946408,0.0002249265,0.02082615,0.01736717,0.004538511,0.01429261,0.9257371],"study_design_scores_gemma":[0.0001160832,0.0002801112,0.005052134,0.00005567854,0.0001873411,0.001439285,0.0001957757,0.9434766,0.015534,0.0214714,0.01215787,0.00003369222],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1037882,0.001170718,0.8797337,0.000327609,0.000162385,0.0004484724,0.003385322,0.008791653,0.002192128],"genre_scores_gemma":[0.2790891,0.0004487055,0.7112548,0.00009837242,0.00007398587,0.0003044787,0.006298919,0.0002422971,0.002189409],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003480332,"threshold_uncertainty_score":0.00626862,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4387703477","doi":"10.1007/s10618-023-00970-4","title":"Mondrian forest for data stream classification under memory constraints","year":2023,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Foundation for Innovation","keywords":"Mondrian; Computer science; Trimming; Data stream mining; Node (physics); Data stream; Data mining; Theoretical computer science; Machine learning; Programming language","authors":[{"name":"Martin Khannouz","is_ca":true},{"name":"Tristan Glatard","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1783262536153769,"gpt":0.3657060325522407,"spread":0.1873797789368638,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001800633,0.0005910813,0.001588384,0.001546335,0.001089959,0.001672622,0.0023596,0.001176492,0.004860964],"category_scores_gemma":[0.009726363,0.0005369285,0.001017628,0.002088959,0.0006637584,0.003505663,0.001373555,0.002054146,0.001189828],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008453457,"about_ca_system_score_gemma":0.00186222,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007886954,"about_ca_topic_score_gemma":0.01266271,"domain_scores_codex":[0.9992234,0.0001925621,0.00007435077,0.0001820676,0.0002180889,0.0001095386],"domain_scores_gemma":[0.9974898,0.001299028,0.0001453263,0.0005871967,0.0003658051,0.000112857],"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":[0.0009173292,0.0002592109,0.004142168,0.0004013326,0.0001602047,0.0003565581,0.0001646854,0.1692622,0.002823952,0.1283385,0.03451343,0.6586605],"study_design_scores_gemma":[0.00003206641,0.00003159571,0.0003795568,0.00003358449,0.00002377548,0.0001004342,0.00002463617,0.8728952,0.0006574555,0.1221481,0.003662273,0.00001136922],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03462319,0.002969654,0.9520457,0.001268012,0.0003037395,0.0001186218,0.001301603,0.003210745,0.004158695],"genre_scores_gemma":[0.3980083,0.001882538,0.5815732,0.0005465919,0.0005836469,0.0003822458,0.003936749,0.0005300047,0.01255677],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007886954,"threshold_uncertainty_score":0.01626158,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408175168","doi":"10.1007/s10618-024-01084-1","title":"Efficient outlier detection in numerical and categorical data","year":2025,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":2,"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":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior; Fundação de Amparo à Pesquisa do Estado de São Paulo; Carnegie Mellon University","keywords":"Categorical variable; Anomaly detection; Outlier; Computer science; Data mining; Artificial intelligence; Pattern recognition (psychology); Machine learning","authors":[{"name":"Eugênio F. Cabral","is_ca":false},{"name":"Braulio V. Sánchez Vinces","is_ca":false},{"name":"Guilherme D. F. Silva","is_ca":false},{"name":"Jörg Sander","is_ca":true},{"name":"Robson L. F. Cordeiro","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03590234063653296,"gpt":0.3165941644719431,"spread":0.2806918238354101,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002319105,0.0008010081,0.001732998,0.003314694,0.0005712726,0.001835923,0.002182402,0.001029799,0.001025085],"category_scores_gemma":[0.01265806,0.0003461011,0.001033088,0.003601987,0.0007336323,0.001816557,0.001733969,0.001639634,0.0009160505],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007080211,"about_ca_system_score_gemma":0.001090802,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002984393,"about_ca_topic_score_gemma":0.003051393,"domain_scores_codex":[0.9965414,0.0007152933,0.0002886851,0.0008199353,0.001378929,0.0002557964],"domain_scores_gemma":[0.9915295,0.003505219,0.001140232,0.001207139,0.002256103,0.0003617849],"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":[0.0007788211,0.0004618866,0.05017883,0.0008298702,0.0004818234,0.0006470408,0.0003748667,0.189986,0.03207624,0.01410832,0.02496569,0.6851107],"study_design_scores_gemma":[0.00002218203,0.00007385336,0.003295305,0.00002987154,0.00002239073,0.0002696,0.0001757893,0.9690149,0.00919436,0.01431685,0.003555957,0.00002889099],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1038752,0.001151969,0.884253,0.0008388481,0.0002338018,0.0001195896,0.001508223,0.006576526,0.001442814],"genre_scores_gemma":[0.525085,0.0003125083,0.4689063,0.0001794079,0.0001327879,0.0001018397,0.003634019,0.0002408339,0.001407339],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003314694,"threshold_uncertainty_score":0.01226473,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4391572936","doi":"10.1007/s10618-023-01001-y","title":"VEM$$^2$$L: an easy but effective framework for fusing text and structure knowledge on sparse knowledge graph completion","year":2024,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Topic Modeling","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"National Natural Science Foundation of China","keywords":"Computer science; Knowledge graph; Graph; Information retrieval; Artificial intelligence; Theoretical computer science","authors":[{"name":"Tao He","is_ca":false},{"name":"Ming Liu","is_ca":false},{"name":"Yixin Cao","is_ca":false},{"name":"Meng Qu","is_ca":true},{"name":"Zihao Zheng","is_ca":false},{"name":"Bing Qin","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05877690887851249,"gpt":0.3329948927339794,"spread":0.2742179838554669,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00280218,0.001083771,0.001531269,0.002935064,0.001053445,0.002067932,0.004154458,0.002150724,0.009457109],"category_scores_gemma":[0.01445551,0.0007808197,0.001775818,0.00223671,0.001091374,0.004477466,0.005851178,0.00369725,0.003232017],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001035915,"about_ca_system_score_gemma":0.0019655,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007487746,"about_ca_topic_score_gemma":0.01648667,"domain_scores_codex":[0.9977039,0.0008728211,0.000112038,0.00049551,0.0006216493,0.0001940812],"domain_scores_gemma":[0.9953257,0.002365846,0.0001862434,0.001178382,0.0006750773,0.000268701],"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":[0.0003550678,0.000385514,0.001166164,0.000431916,0.0001461326,0.0002150114,0.0002872201,0.07831257,0.005861424,0.1223718,0.05613363,0.7343336],"study_design_scores_gemma":[0.00002450296,0.00004156484,0.0001356405,0.00003083565,0.00001585878,0.00006237259,0.00004069724,0.8822445,0.002031197,0.1061703,0.009178226,0.00002426297],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001311349,0.00009451254,0.9948121,0.0002020745,0.00004806249,0.00006950744,0.0003700612,0.002507496,0.000584938],"genre_scores_gemma":[0.06090938,0.0001693587,0.9314928,0.0003244801,0.0001258373,0.0002848813,0.002043371,0.0006495243,0.004000296],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009457109,"threshold_uncertainty_score":0.03163719,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2091837657","doi":"10.1007/s10618-010-0199-1","title":"Improving constrained pattern mining with first-fail-based heuristics","year":2010,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"HEC Montréal; Polytechnique Montréal; École de Technologie Supérieure","funders":"","keywords":"Heuristics; Data mining; Computer science; Process (computing); Data stream mining; State (computer science); Sequential Pattern Mining; Machine learning; Artificial intelligence; Algorithm","authors":[{"name":"Christian Desrosiers","is_ca":true},{"name":"Philippe Galinier","is_ca":true},{"name":"Alain Hertz","is_ca":true},{"name":"Pierre Hansen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02107824697066401,"gpt":0.2557087373958644,"spread":0.2346304904252004,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005993431,0.002234811,0.004463254,0.004747842,0.001699416,0.003296301,0.004666917,0.002602617,0.005735678],"category_scores_gemma":[0.03783927,0.001323337,0.002740656,0.004623188,0.001501889,0.005825853,0.003271029,0.003460167,0.001535822],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001001892,"about_ca_system_score_gemma":0.004161924,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00839112,"about_ca_topic_score_gemma":0.01404892,"domain_scores_codex":[0.9944722,0.00196791,0.0006090587,0.0009725585,0.001467687,0.0005105138],"domain_scores_gemma":[0.9622105,0.02729852,0.001278718,0.005198038,0.003333235,0.0006810351],"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":[0.002134752,0.001328364,0.01408355,0.000910112,0.001037756,0.0008339862,0.0004028815,0.2418978,0.005191872,0.02248284,0.01971238,0.6899837],"study_design_scores_gemma":[0.0001935503,0.0002473155,0.000850538,0.00005622533,0.0002303498,0.000311142,0.00009199604,0.9667842,0.00220951,0.02673181,0.002255814,0.00003742779],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04539374,0.001567837,0.9453356,0.00067141,0.000229465,0.0004083736,0.0005473457,0.003188808,0.002657472],"genre_scores_gemma":[0.2924267,0.0004775158,0.7002919,0.0009168333,0.000194783,0.0003770455,0.001914073,0.0005453418,0.002855823],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00839112,"threshold_uncertainty_score":0.03169662,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4414033870","doi":"10.1007/s10618-025-01158-8","title":"Detecting and reacting to smart home novelties","year":2025,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Network for Business Sustainability","funders":"Defense Advanced Research Projects Agency; National Institutes of Health; National Science Foundation","keywords":"Computer science","authors":[{"name":"Lawrence B. Holder","is_ca":false},{"name":"Baxter Eaves","is_ca":true},{"name":"Patrick Shafto","is_ca":true},{"name":"Christopher Pereyda","is_ca":false},{"name":"Brian Thomas","is_ca":false},{"name":"Diane J. Cook","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03636783725970738,"gpt":0.3125553640990832,"spread":0.2761875268393759,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001827295,0.0004691359,0.0006719531,0.001521355,0.0004244852,0.001619337,0.0009692052,0.001162471,0.001328945],"category_scores_gemma":[0.01500219,0.0002179034,0.0003986345,0.0009054086,0.0005155649,0.002341795,0.001239688,0.001383028,0.0004563744],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004383357,"about_ca_system_score_gemma":0.0005154462,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008277817,"about_ca_topic_score_gemma":0.00129791,"domain_scores_codex":[0.9984189,0.0002649691,0.00008595194,0.0003717886,0.0007182105,0.0001402306],"domain_scores_gemma":[0.9908338,0.005010906,0.00173091,0.000782361,0.001190269,0.0004518194],"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":[0.003021345,0.001290423,0.3893949,0.0006889558,0.0007001551,0.008268006,0.003368596,0.03435168,0.05789832,0.01723228,0.02404943,0.4597358],"study_design_scores_gemma":[0.0001171015,0.0008138959,0.08749126,0.0001165793,0.0003965943,0.003388873,0.003692097,0.8188317,0.03437054,0.02863334,0.0220373,0.0001107537],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9031714,0.0009449672,0.0804425,0.002616358,0.0007185524,0.0002379846,0.001246953,0.001683685,0.008937576],"genre_scores_gemma":[0.9786916,0.0002697557,0.01823789,0.0002871464,0.0001572775,0.00004104007,0.0005345925,0.00004608177,0.001734571],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001827295,"threshold_uncertainty_score":0.00966382,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2093446553","doi":"10.1007/s10618-005-0022-6","title":"An Erratum on #x201C;Pushing Convertible Constraints in Frequent Itemset Mining#x201D;","year":2006,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia; Simon Fraser University","funders":"","keywords":"Prefix; Order (exchange); Combinatorics; Convertible; Function (biology); Mathematics; Physics; Philosophy; Linguistics; Biology; Business; Genetics","authors":[{"name":"Jian Pei","is_ca":true},{"name":"Jiawei Han","is_ca":false},{"name":"Laks V. S. Lakshmanan","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03631538169184479,"gpt":0.2998001775418355,"spread":0.2634847958499907,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004950527,0.002240018,0.002274837,0.005465471,0.007138385,0.006591117,0.004811122,0.01028747,0.02493144],"category_scores_gemma":[0.06169618,0.001487345,0.001791573,0.004946713,0.003327701,0.004014779,0.002448656,0.0133175,0.01644577],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00517638,"about_ca_system_score_gemma":0.006817111,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.04665974,"about_ca_topic_score_gemma":0.06422651,"domain_scores_codex":[0.9915079,0.00126125,0.001069066,0.0008664301,0.004780264,0.0005151304],"domain_scores_gemma":[0.9519758,0.01721399,0.00155588,0.00327589,0.02407401,0.001904448],"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":[0.00002311175,0.00001185824,0.00004948417,0.00005392134,0.000007245251,0.00008787886,0.00001474853,0.00006759068,0.00008394413,0.001303857,0.9921515,0.006144894],"study_design_scores_gemma":[0.00004484392,0.00003574546,0.0006525701,0.0001999083,0.0000529155,0.0003324395,0.00008069881,0.001660678,0.0008048388,0.006812119,0.9892615,0.00006176333],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"other","genre_scores_codex":[0.0004652782,0.002023428,0.008980656,0.1671294,0.8091916,0.00009499825,0.001627104,0.0008451871,0.00964231],"genre_scores_gemma":[0.01272897,0.008806128,0.05266488,0.2560089,0.4477673,0.0004989941,0.004959193,0.001938171,0.2146274],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.04665974,"threshold_uncertainty_score":0.0927763,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4401666276","doi":"10.1007/s10618-024-01054-7","title":"Bayesian network Motifs for reasoning over heterogeneous unlinked datasets","year":2024,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada Research Chairs; University of Toronto; University of New Brunswick","funders":"","keywords":"Computer science; Probabilistic logic; Bayesian network; Python (programming language); Data mining; Aggregate (composite); Real world data; Bayesian probability; Machine learning; Set (abstract data type); Artificial intelligence; Data science","authors":[{"name":"Yi Sui","is_ca":true},{"name":"Alex C. Kwan","is_ca":true},{"name":"Alexander W Olson","is_ca":true},{"name":"Scott Sanner","is_ca":true},{"name":"Daniel Aaron Silver","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1674946705782729,"gpt":0.4358667777938899,"spread":0.2683721072156171,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006303884,0.0009469164,0.001724678,0.005379261,0.001370603,0.003377518,0.003754775,0.003106374,0.005181895],"category_scores_gemma":[0.06944677,0.001202293,0.002209681,0.004837027,0.001509382,0.01009097,0.00346355,0.003721661,0.0009479563],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002118853,"about_ca_system_score_gemma":0.002374718,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009635517,"about_ca_topic_score_gemma":0.01450601,"domain_scores_codex":[0.9952177,0.001625648,0.0004265507,0.001037123,0.001462882,0.0002300921],"domain_scores_gemma":[0.9605373,0.03230933,0.002391583,0.002392633,0.001678909,0.0006903972],"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":[0.0006342022,0.0004632248,0.01036486,0.001027713,0.0007022623,0.001006005,0.0008345693,0.3029706,0.003286767,0.3915555,0.00920718,0.2779472],"study_design_scores_gemma":[0.00003602022,0.00002429004,0.0003302174,0.00007642182,0.00007483584,0.000129336,0.00006295311,0.6005837,0.000652386,0.3958277,0.002185132,0.00001702007],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01265743,0.0005827165,0.9825656,0.000777947,0.00004645783,0.000131952,0.001038169,0.001276814,0.0009229041],"genre_scores_gemma":[0.270836,0.000897358,0.7210087,0.0004082642,0.0001431424,0.0004469147,0.003841622,0.0002950105,0.002122839],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009635517,"threshold_uncertainty_score":0.03333849,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3097249106","doi":"10.1007/s10618-020-00720-w","title":"An exemplar-based clustering using efficient variational message passing","year":2020,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec en Outaouais","funders":"","keywords":"Cluster analysis; Computer science; Correlation clustering; Message passing; Affinity propagation; Constrained clustering; Graph partition; CURE data clustering algorithm; Canopy clustering algorithm; Data mining; Graph; Theoretical computer science; Artificial intelligence","authors":[{"name":"Mohamed Hamza Ibrahim","is_ca":true},{"name":"Rokia Missaoui","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1118581280584065,"gpt":0.3593757391597357,"spread":0.2475176111013292,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002456837,0.001011569,0.00277192,0.001669681,0.001626472,0.001943945,0.005148914,0.002975965,0.003393171],"category_scores_gemma":[0.005772321,0.0012358,0.001820158,0.00223619,0.001632239,0.002377099,0.003946987,0.002320094,0.001389056],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001693037,"about_ca_system_score_gemma":0.002405277,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01212762,"about_ca_topic_score_gemma":0.01014228,"domain_scores_codex":[0.9984395,0.0004557483,0.00008758149,0.0003377577,0.0005202307,0.0001591483],"domain_scores_gemma":[0.9978253,0.0009211592,0.0001286929,0.0003438941,0.0006233511,0.0001576904],"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":[0.0002172192,0.0001754109,0.0007604932,0.0001576486,0.0001849709,0.00009706475,0.0002502372,0.7117519,0.01041448,0.05946484,0.003875877,0.2126497],"study_design_scores_gemma":[0.000006555282,0.00001462823,0.00003655343,0.000002669619,0.000007160662,0.00001692642,0.000007862978,0.9942445,0.0008175471,0.004456814,0.0003811786,0.000007607154],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003207884,0.0000544327,0.9959266,0.00007992281,0.00003114687,0.00003334737,0.00002178038,0.0002453365,0.000399624],"genre_scores_gemma":[0.1241425,0.0001354405,0.8710372,0.0001414105,0.00007982337,0.0001911057,0.0003012528,0.0002223605,0.003748966],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01212762,"threshold_uncertainty_score":0.02411401,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2810490377","doi":"10.1007/s10618-020-00677-w","title":"Model-based exception mining for object-relational data","year":2020,"lang":"en","type":"preprint","venue":"Data Mining and Knowledge Discovery","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"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","keywords":"Computer science; Object (grammar); Outlier; Data mining; Metric (unit); Bayesian network; Probabilistic logic; Relational model; Representation (politics); Relational database; Population; Statistical relational learning; Artificial intelligence; Object model; Graphical model; Data modeling; Statistical model; Database","authors":[{"name":"Fatemeh Riahi","is_ca":true},{"name":"Oliver Schulte","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2050746449930389,"gpt":0.3555595803813095,"spread":0.1504849353882706,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004352093,0.0009239136,0.002081575,0.003184909,0.001252931,0.004502757,0.00340274,0.001231738,0.001172003],"category_scores_gemma":[0.02066854,0.0006942317,0.002827458,0.003433804,0.001138874,0.004864201,0.002699333,0.00227272,0.0006219551],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008890066,"about_ca_system_score_gemma":0.002615137,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004516327,"about_ca_topic_score_gemma":0.007145752,"domain_scores_codex":[0.9947812,0.001141225,0.0006910452,0.0009095572,0.002067462,0.0004095623],"domain_scores_gemma":[0.9861572,0.005290629,0.001381165,0.004974446,0.00182412,0.00037241],"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":[0.001143259,0.0007880978,0.04325555,0.0008656931,0.0009134103,0.001157801,0.0007056828,0.2692659,0.01156313,0.07025283,0.01751538,0.5825733],"study_design_scores_gemma":[0.00003735269,0.00009593097,0.001055875,0.00005068728,0.0001473558,0.0003921858,0.0001145043,0.9068093,0.007525007,0.08075069,0.002988164,0.00003290628],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02940197,0.0004378566,0.9637769,0.0003750746,0.0001154815,0.0001544809,0.0007610053,0.004125202,0.000852012],"genre_scores_gemma":[0.4028634,0.000396321,0.5919204,0.000245421,0.000094709,0.0001669089,0.002601543,0.0004353208,0.001275992],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004516327,"threshold_uncertainty_score":0.02301633,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4417267046","doi":"10.1007/s10618-025-01180-w","title":"Certifying robustness of graph convolutional networks for node perturbation with polyhedra abstract interpretation","year":2025,"lang":"en","type":"article","venue":"Data Mining and Knowledge Discovery","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"Fonds de recherche du Québec – Nature et technologies; Nemzeti Kutatási Fejlesztési és Innovációs Hivatal; Mitacs; Knut och Alice Wallenbergs Stiftelse","keywords":"Robustness (evolution); Graph; Convolutional neural network; Adversarial system; Polyhedron; Training set","authors":[{"name":"Boqi Chen","is_ca":true},{"name":"Kristóf Marussy","is_ca":false},{"name":"Oszkár Semeráth","is_ca":false},{"name":"Gunter Mussbacher","is_ca":true},{"name":"Dániel Varró","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03197076562914049,"gpt":0.2958814579966345,"spread":0.263910692367494,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004447351,0.001502866,0.001682944,0.002016071,0.0008619669,0.002860314,0.003617936,0.002804296,0.003980623],"category_scores_gemma":[0.04567223,0.0007803928,0.00199867,0.001131194,0.004147538,0.005383641,0.005294321,0.004495289,0.0005867841],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002671786,"about_ca_system_score_gemma":0.001719799,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004125218,"about_ca_topic_score_gemma":0.003024269,"domain_scores_codex":[0.9960362,0.0009125374,0.0002245861,0.001318504,0.001108008,0.0004002006],"domain_scores_gemma":[0.9753036,0.01531375,0.001831755,0.005014525,0.001963158,0.000573186],"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":[0.0006589548,0.00008332142,0.003397385,0.000299231,0.0002563552,0.0003751158,0.0001680826,0.7384124,0.007003783,0.1924198,0.00511218,0.05181354],"study_design_scores_gemma":[0.00001149241,0.00002769238,0.0002884245,0.00001876611,0.00001826079,0.0000372759,0.00001570373,0.8305683,0.0022788,0.1663998,0.0003245323,0.00001104777],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08703447,0.0003745577,0.9038057,0.001303002,0.0001777101,0.00008572472,0.0007600314,0.002387424,0.004071386],"genre_scores_gemma":[0.9225166,0.0003207202,0.07161448,0.0004674892,0.0002165443,0.0001212186,0.001573608,0.0005769086,0.002592405],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004447351,"threshold_uncertainty_score":0.02352017,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}