{"meta":{"query_hash":"b633b755ce29","filters":{"venue":"Big Data and Information Analytics"},"cohort_total":18,"direct_labels_cover":0,"predictions_cover":18,"exported":18,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/b633b755ce29","api":"https://metacan.xera.ac/api/v1/cohort?venue=Big+Data+and+Information+Analytics"},"results":[{"id":"W2476626105","doi":"10.3934/bdia.2016.1.1","title":"ACO-based solution for computation offloading in mobile cloud computing","year":2015,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"IoT and Edge/Fog Computing","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"Natural Science Foundation of Hunan Province; National Natural Science Foundation of China","keywords":"Mobile cloud computing; Computation offloading; Cloud computing; Computer science; Ant colony optimization algorithms; Distributed computing; Computation; Mobile device; Mobile computing; Computer network; Edge computing; Operating system; Algorithm","score_opus":0.12292265115255017,"score_gpt":0.31653533143521567,"score_spread":0.19361268028266548,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2476626105","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.031469602,0.0010501306,0.9483518,0.0006751517,0.00030593903,0.00015058897,0.000093157105,0.00036605037,0.017537491],"genre_scores_gemma":[0.7211941,0.0007941138,0.26870525,0.0002994817,0.00014913746,0.00022970568,0.00015119562,0.00012622197,0.008350816],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996475,0.000089665504,0.000018884368,0.000060038896,0.00010297362,0.00008097799],"domain_scores_gemma":[0.99964833,0.00016001533,0.00004143756,0.000026562993,0.00009020231,0.000033459564],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003564974,0.0006992975,0.0009742856,0.0006080797,0.00067748607,0.00083495857,0.001126839,0.00093148777,0.0022107153],"category_scores_gemma":[0.0010236152,0.00030238242,0.00068997213,0.0009201891,0.0004588022,0.00052167283,0.00072195556,0.0007526442,0.00032981738],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00008067266,0.00006176392,0.0005685216,0.00011026204,0.000057938232,0.00019161483,0.000041086863,0.9309894,0.0028290425,0.015275533,0.0031563141,0.046637956],"study_design_scores_gemma":[0.000010194154,0.000010298544,0.00007218208,0.00000402237,0.000004778633,0.00002318713,0.0000074363984,0.99727255,0.00021176784,0.0016856005,0.0006948873,0.0000031249267],"about_ca_topic_score_codex":0.007893567,"about_ca_topic_score_gemma":0.007925297,"teacher_disagreement_score":0.007893567,"about_ca_system_score_codex":0.00048475363,"about_ca_system_score_gemma":0.0012667044,"threshold_uncertainty_score":0.015695214},"labels":[],"label_agreement":null},{"id":"W2497857151","doi":"10.3934/bdia.2016.1.31","title":"What's the big deal about big data?","year":2015,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Big data; Variety (cybernetics); Data science; Theme (computing); Novelty; Computer science; Analytics; Scale (ratio); Informatics; Business intelligence; Data analysis; Cloud computing; Knowledge management; World Wide Web; Engineering; Artificial intelligence; Data mining; Psychology","score_opus":0.502372524389615,"score_gpt":0.40834206601971434,"score_spread":0.09403045836990065,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2497857151","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0015331197,0.07273566,0.008975883,0.89584994,0.013573533,0.000046299472,0.00029426522,0.000120921904,0.00687034],"genre_scores_gemma":[0.16039552,0.2769211,0.040847592,0.40867546,0.09762766,0.0004865957,0.0011559408,0.0005971329,0.013293057],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98027426,0.009686267,0.0006940655,0.0016983537,0.0062508835,0.0013962825],"domain_scores_gemma":[0.9046179,0.07173311,0.003581856,0.0048229764,0.008565051,0.006678973],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.031155156,0.001010451,0.0018871345,0.0036800124,0.0064878208,0.025789535,0.0030189094,0.008401311,0.007082895],"category_scores_gemma":[0.060389813,0.00077637663,0.0015416486,0.006310258,0.01514666,0.03921841,0.0058623725,0.013699873,0.002680054],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00023834402,0.00014281369,0.007084598,0.0036515696,0.0003997665,0.00037895128,0.0028963753,0.001805718,0.00053323375,0.30188468,0.48594922,0.19503477],"study_design_scores_gemma":[0.000035026074,0.00006210655,0.001967863,0.0030642976,0.000071725735,0.00031298172,0.0062507074,0.0016960109,0.000360773,0.5567345,0.42934725,0.00009669064],"about_ca_topic_score_codex":0.0041067027,"about_ca_topic_score_gemma":0.0046219174,"teacher_disagreement_score":0.031155156,"about_ca_system_score_codex":0.004888636,"about_ca_system_score_gemma":0.009510067,"threshold_uncertainty_score":0.16476619},"labels":[],"label_agreement":null},{"id":"W2527216788","doi":"10.3934/bdia.2016006","title":"Detecting coalition attacks in online advertising: A hybrid data mining approach","year":2016,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Trent University","funders":"Mitacs","keywords":"Computer science; Computer security; Data mining","score_opus":0.12153644611219627,"score_gpt":0.2952623683958516,"score_spread":0.17372592228365535,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2527216788","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.21301652,0.00080880325,0.78004825,0.0011043797,0.0000533316,0.0004121483,0.0007393835,0.0011802956,0.0026369316],"genre_scores_gemma":[0.76743,0.000180491,0.23073113,0.00014588588,0.000065023305,0.00014139393,0.00067389134,0.000036553887,0.00059556344],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9933021,0.0026669335,0.0006829312,0.0010510167,0.001908719,0.0003883964],"domain_scores_gemma":[0.9867754,0.006874593,0.0018018822,0.0018116492,0.0021881605,0.0005482286],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008353772,0.0013318274,0.00209476,0.012246212,0.0010021745,0.0042674523,0.0022985425,0.0015937984,0.00038177127],"category_scores_gemma":[0.01247383,0.0003727676,0.0015867935,0.0063280673,0.0010240336,0.004599541,0.0028138435,0.0016236122,0.00026096829],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00070454064,0.002180619,0.2391811,0.00040652213,0.0016418785,0.0006846927,0.0011595353,0.18184976,0.011065081,0.031212885,0.003811736,0.52610165],"study_design_scores_gemma":[0.000013877321,0.00011967832,0.009119771,0.000022993952,0.00007352853,0.00022331193,0.00027959602,0.9727331,0.002879987,0.013361966,0.0011321168,0.000040126877],"about_ca_topic_score_codex":0.0030694604,"about_ca_topic_score_gemma":0.0033038838,"teacher_disagreement_score":0.012246212,"about_ca_system_score_codex":0.0011727934,"about_ca_system_score_gemma":0.0011240548,"threshold_uncertainty_score":0.0441795},"labels":[],"label_agreement":null},{"id":"W2552308941","doi":"10.3934/bdia.2016008","title":"Time aware topic based recommender system","year":2016,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Recommender system; Computer science; Collaborative filtering; Information retrieval; Filter (signal processing); Cold start (automotive); World Wide Web; Topic model","score_opus":0.0647731355806819,"score_gpt":0.2560971786824018,"score_spread":0.19132404310171988,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2552308941","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.038660884,0.0073436378,0.92965585,0.001301519,0.00096070813,0.00045932856,0.0019221661,0.0046631503,0.015032829],"genre_scores_gemma":[0.5819437,0.008769489,0.3606582,0.00075123523,0.0013781547,0.0006762294,0.0047285,0.0003195735,0.04077493],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.998431,0.00032235644,0.00013892012,0.00047711213,0.0004791554,0.00015137848],"domain_scores_gemma":[0.9972536,0.0008513284,0.00021076207,0.00036955113,0.0011801593,0.00013466329],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014877624,0.0010589629,0.002098077,0.0017880421,0.0012985367,0.0021047315,0.0021733649,0.0019975624,0.0047914255],"category_scores_gemma":[0.0042333277,0.00055730663,0.0015198895,0.0022552651,0.00029865323,0.002612927,0.0009254276,0.0015095538,0.0052909246],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009919099,0.00064461667,0.016469661,0.0012670009,0.0010557707,0.0011691328,0.001033508,0.13247953,0.046325043,0.020113314,0.059792984,0.7186575],"study_design_scores_gemma":[0.00015393144,0.0003436626,0.0043345196,0.0000728306,0.0007222425,0.0013219279,0.0002709236,0.9173769,0.009583982,0.00997843,0.055642202,0.00019850401],"about_ca_topic_score_codex":0.012038524,"about_ca_topic_score_gemma":0.013488457,"teacher_disagreement_score":0.012038524,"about_ca_system_score_codex":0.0005819452,"about_ca_system_score_gemma":0.0012950641,"threshold_uncertainty_score":0.023936927},"labels":[],"label_agreement":null},{"id":"W2553369270","doi":"10.3934/bdia.2016009","title":"Manifold data mining helps businesses grow more effectively","year":2016,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Manifold (fluid mechanics); Nonlinear dimensionality reduction; Data mining; Data science; Business; Computer science; Engineering; Artificial intelligence; Mechanical engineering","score_opus":0.08789576314961048,"score_gpt":0.2890048371131446,"score_spread":0.20110907396353414,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2553369270","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.09710267,0.015671682,0.39616182,0.2523509,0.003940197,0.0006571058,0.007662142,0.0143164685,0.21213703],"genre_scores_gemma":[0.47713196,0.015889365,0.4512689,0.0077293147,0.0028903103,0.00040518166,0.00786041,0.0020572487,0.03476731],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99531186,0.001072336,0.0002322595,0.00069113687,0.0022844514,0.00040788672],"domain_scores_gemma":[0.9837841,0.0043143174,0.0012900278,0.0044850255,0.0044479775,0.0016785566],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0070562907,0.0014047741,0.0016123248,0.0052685263,0.0038195492,0.015675012,0.0026010566,0.0021349275,0.010802213],"category_scores_gemma":[0.03297362,0.00079511217,0.0014407666,0.00973438,0.003591967,0.02868064,0.01090026,0.0050454116,0.010592469],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017227986,0.00026537132,0.017097525,0.00044144274,0.00019852885,0.00023499028,0.0020103264,0.007770222,0.0013666274,0.24380837,0.29430535,0.43232897],"study_design_scores_gemma":[0.000038890863,0.000081139435,0.0029426091,0.00029692284,0.00006761239,0.0001524477,0.0043465244,0.031037316,0.0019222508,0.5138356,0.4451996,0.000079022626],"about_ca_topic_score_codex":0.0061003193,"about_ca_topic_score_gemma":0.011297418,"teacher_disagreement_score":0.015675012,"about_ca_system_score_codex":0.0018077305,"about_ca_system_score_gemma":0.0052041914,"threshold_uncertainty_score":0.037317693},"labels":[],"label_agreement":null},{"id":"W2604624992","doi":"10.3934/bdia.2016011","title":"Analyzing opinion dynamics in online social networks","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Opinion Dynamics and Social Influence","field":"Physics and Astronomy","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Skepticism; Empathy; Dynamics (music); Social network (sociolinguistics); Population; Focus (optics); Social media; Psychology; Social psychology; Computer science; Data science; Sociology; World Wide Web; Epistemology","score_opus":0.054415160693912556,"score_gpt":0.32772892051200514,"score_spread":0.2733137598180926,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2604624992","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.70016426,0.00075137994,0.29113814,0.0017605961,0.00006949705,0.00009409239,0.0005567965,0.00016145995,0.0053036767],"genre_scores_gemma":[0.9876252,0.00029674632,0.010826295,0.00007434147,0.00007628325,0.000060322873,0.00019180516,0.000016118982,0.00083282444],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987508,0.0005614212,0.000045059223,0.0002473936,0.00026429817,0.00013107863],"domain_scores_gemma":[0.98659766,0.010161197,0.0018362001,0.00058007066,0.0005574265,0.0002674522],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021056621,0.00041079157,0.000599317,0.0014198885,0.0005413136,0.0012614682,0.0008471071,0.0009842404,0.0014619282],"category_scores_gemma":[0.020048149,0.00022355445,0.0005498187,0.0010412531,0.0008530206,0.0032727134,0.00089377305,0.0007932335,0.0002684447],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002821907,0.00033138003,0.12236211,0.0003980183,0.00041942546,0.0009991637,0.0031745655,0.56149,0.011942484,0.20815739,0.003558102,0.08688514],"study_design_scores_gemma":[0.0000121702205,0.000040749528,0.011929327,0.000019096735,0.000020810918,0.00009488058,0.00048611878,0.9294488,0.0008036465,0.055286847,0.0018364813,0.000021045711],"about_ca_topic_score_codex":0.004166169,"about_ca_topic_score_gemma":0.0022722322,"teacher_disagreement_score":0.004166169,"about_ca_system_score_codex":0.0008468085,"about_ca_system_score_gemma":0.00031192778,"threshold_uncertainty_score":0.011135995},"labels":[],"label_agreement":null},{"id":"W2604652933","doi":"10.3934/bdia.2016012","title":"Modeling daily guest count prediction","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Customer churn and segmentation","field":"Business, Management and Accounting","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Lasso (programming language); Computer science; Data mining; Preprocessor; Data pre-processing; Count data; Transaction data; Feature (linguistics); Data modeling; Regression; Artificial intelligence; Predictive modelling; Machine learning; Regression analysis; Poisson distribution; Database transaction; Statistics; Mathematics; Database","score_opus":0.10183468277054568,"score_gpt":0.2750029553143635,"score_spread":0.17316827254381778,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2604652933","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.088314585,0.0005118177,0.9062373,0.00052703015,0.00012213818,0.000041017778,0.001281037,0.0009547944,0.0020103103],"genre_scores_gemma":[0.8439324,0.00062654173,0.14403927,0.00017696404,0.00035680857,0.00018338666,0.0042658625,0.00016868369,0.00625003],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999426,0.00013762197,0.000028306264,0.00020361121,0.00012647343,0.000077892095],"domain_scores_gemma":[0.99875283,0.0006571541,0.00023234187,0.00011772481,0.00018264657,0.00005735419],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00124649,0.0007940983,0.00096110255,0.00096104527,0.00033012472,0.0009897839,0.0014460303,0.0007946407,0.0015464755],"category_scores_gemma":[0.0034320024,0.00047847192,0.0009374512,0.0014522881,0.00032802808,0.0010998002,0.0006386778,0.0012543555,0.0005182328],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000114964336,0.000118008575,0.01474305,0.00007617179,0.00013019961,0.00015408976,0.00008703473,0.9041814,0.0018579428,0.00869085,0.0040449235,0.06580139],"study_design_scores_gemma":[0.0000010153968,0.000005446134,0.00037526307,0.0000015082634,0.0000028991858,0.000006935326,0.0000036518238,0.9983753,0.00009377005,0.00088947976,0.00024221481,0.0000023917657],"about_ca_topic_score_codex":0.012400194,"about_ca_topic_score_gemma":0.012453394,"teacher_disagreement_score":0.012400194,"about_ca_system_score_codex":0.0005728462,"about_ca_system_score_gemma":0.00081364065,"threshold_uncertainty_score":0.024655998},"labels":[],"label_agreement":null},{"id":"W2611392997","doi":"10.3934/bdia.2016013","title":"A testbed to enable comparisons between competing approaches for computational social choice","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Logic, Reasoning, and Knowledge","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Testbed; Computer science; Rank (graph theory); Context (archaeology); Domain (mathematical analysis); Preference; Voting; Action (physics); Social choice theory; Order (exchange); Field (mathematics); Group decision-making; Artificial intelligence; Operations research; Data science; Machine learning; Engineering","score_opus":0.26739060138568005,"score_gpt":0.3367448301259789,"score_spread":0.06935422874029884,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2611392997","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.3489812,0.0011954487,0.583236,0.005842229,0.00086835446,0.003285212,0.010289048,0.0028959427,0.043406587],"genre_scores_gemma":[0.69878954,0.0004664186,0.2907859,0.0003632129,0.00008670679,0.0024270408,0.004585171,0.0001580266,0.0023378392],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9944253,0.004252242,0.00025475668,0.00032285965,0.0005522622,0.00019256449],"domain_scores_gemma":[0.9788163,0.014622742,0.00072774856,0.003870263,0.0011144965,0.0008484846],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006235409,0.00046615902,0.0005614544,0.0011425137,0.0012251415,0.0016684685,0.0019116548,0.0015197846,0.004798247],"category_scores_gemma":[0.020141479,0.00030894982,0.0005816125,0.0020012418,0.002241863,0.0027303956,0.0018826938,0.0014710628,0.0008022747],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025161298,0.0043345513,0.016819963,0.0012701947,0.00030383086,0.0005922859,0.0019640578,0.379617,0.010100332,0.4577966,0.03156177,0.09312331],"study_design_scores_gemma":[0.0007241297,0.0021926234,0.006272184,0.00018835672,0.00007393395,0.00025814606,0.0014847334,0.71190083,0.008392986,0.20911653,0.059287533,0.0001079803],"about_ca_topic_score_codex":0.0029980224,"about_ca_topic_score_gemma":0.00341299,"teacher_disagreement_score":0.006235409,"about_ca_system_score_codex":0.0016875484,"about_ca_system_score_gemma":0.0014634604,"threshold_uncertainty_score":0.03297639},"labels":[],"label_agreement":null},{"id":"W2736357342","doi":"10.3934/bdia.2017001","title":"First steps in the investigation of automated text annotation with pictures","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Annotation; Computer science; Natural language processing; Artificial intelligence; Information retrieval; Computer graphics (images)","score_opus":0.0545208875358242,"score_gpt":0.2972021565263602,"score_spread":0.24268126899053596,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2736357342","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027734999,0.0011644154,0.94813186,0.003058414,0.00020818524,0.002829319,0.00066252745,0.002824201,0.013386085],"genre_scores_gemma":[0.04873119,0.00039703865,0.94509345,0.0003779969,0.00006004079,0.0011578536,0.00058539474,0.00040097014,0.0031960646],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.98866874,0.0060464293,0.0005989442,0.0017251929,0.0024624767,0.000498133],"domain_scores_gemma":[0.97180474,0.014753112,0.000858583,0.0052470393,0.006878838,0.00045771882],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010170407,0.0020348423,0.0012327436,0.004097397,0.0034593246,0.00828172,0.004243892,0.004189819,0.010240392],"category_scores_gemma":[0.03607284,0.0018546802,0.0016327337,0.003281824,0.004661132,0.014387183,0.004433576,0.0049050706,0.0042863293],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011162776,0.0018169776,0.0070400834,0.0040845093,0.00022780107,0.0026155612,0.014441453,0.00993445,0.17747085,0.17090698,0.017382804,0.5929623],"study_design_scores_gemma":[0.00027065616,0.0018980369,0.010781414,0.001351751,0.00023024045,0.0042143515,0.014080438,0.16909021,0.3675706,0.15215147,0.27775195,0.00060892705],"about_ca_topic_score_codex":0.006467499,"about_ca_topic_score_gemma":0.005394255,"teacher_disagreement_score":0.010240392,"about_ca_system_score_codex":0.0027572603,"about_ca_system_score_gemma":0.0036085013,"threshold_uncertainty_score":0.053786874},"labels":[],"label_agreement":null},{"id":"W2739232917","doi":"10.3934/bdia.2017003","title":"Rendering website traffic data into interactive taste graph visualizations","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario College of Art and Design","funders":"","keywords":"Rendering (computer graphics); Computer science; Computer graphics (images); Graph; World Wide Web; Multimedia; Theoretical computer science","score_opus":0.11087586483674504,"score_gpt":0.35835827069580023,"score_spread":0.24748240585905518,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2739232917","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.070943624,0.0002923533,0.80717695,0.0013040137,0.00055817707,0.00052141526,0.023111684,0.075654976,0.02043681],"genre_scores_gemma":[0.3210499,0.00046940858,0.64260405,0.00050497305,0.00021023604,0.0011135421,0.017383145,0.009175162,0.007489609],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99947304,0.00014780072,0.00004434284,0.00009307736,0.00019163369,0.000050235758],"domain_scores_gemma":[0.9954249,0.002662569,0.00018896276,0.0007582429,0.0007785839,0.0001868555],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012187869,0.001010054,0.0005024433,0.0036336586,0.0005358773,0.002953465,0.0007415218,0.00068230194,0.017065078],"category_scores_gemma":[0.007921466,0.0004611573,0.0008179308,0.002130938,0.00042101368,0.0019046403,0.0022803685,0.0013127213,0.0025814627],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0018214666,0.00081469613,0.016419401,0.0024694577,0.00041874818,0.0020288164,0.016158784,0.0443648,0.10169577,0.060786385,0.20636712,0.5466546],"study_design_scores_gemma":[0.0004254012,0.00038150468,0.020532772,0.00049032114,0.00022518719,0.0011151173,0.0052701877,0.38268194,0.066297136,0.13511053,0.38700053,0.00046942753],"about_ca_topic_score_codex":0.0020780624,"about_ca_topic_score_gemma":0.003661222,"teacher_disagreement_score":0.017065078,"about_ca_system_score_codex":0.00034701763,"about_ca_system_score_gemma":0.000545756,"threshold_uncertainty_score":0.057088375},"labels":[],"label_agreement":null},{"id":"W2767704871","doi":"10.3934/bdia.2017015","title":"Identifying electronic gaming machine gambling personae through unsupervised session classification","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Gambling Behavior and Treatments","field":"Psychology","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"","keywords":"Session (web analytics); Outlier; Psychology; Identification (biology); Exploratory research; Applied psychology; Computer science; Artificial intelligence; World Wide Web","score_opus":0.41945866055791536,"score_gpt":0.45244340705503866,"score_spread":0.032984746497123296,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2767704871","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9408794,0.00025533294,0.05163989,0.00022896937,0.000058104586,0.00031123593,0.0036707614,0.00050386495,0.002452433],"genre_scores_gemma":[0.96453464,0.00012259574,0.029607166,0.00005031646,0.000026642965,0.00020481195,0.0041240533,0.000035612335,0.0012940742],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99904114,0.00026532367,0.000103479404,0.00024178441,0.00020382067,0.0001444043],"domain_scores_gemma":[0.9972817,0.0010787391,0.0005486333,0.00035425703,0.00056439097,0.00017227634],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010814283,0.00047133229,0.0005012242,0.002515674,0.00039971902,0.0007926014,0.00066761667,0.0005270736,0.0013913136],"category_scores_gemma":[0.0050275954,0.00013300312,0.0005654796,0.0016071285,0.00022251393,0.0006982503,0.00087396044,0.00061758904,0.0006938345],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00073194853,0.0009436288,0.7403539,0.00030773616,0.0002567101,0.00036500822,0.0018232304,0.005365412,0.006290862,0.0016489273,0.0062523913,0.23566021],"study_design_scores_gemma":[0.00003339334,0.00035202465,0.74265987,0.000152847,0.00016961335,0.0011109792,0.003054026,0.23068264,0.0057757096,0.005892583,0.010029517,0.00008679947],"about_ca_topic_score_codex":0.00443075,"about_ca_topic_score_gemma":0.008092934,"teacher_disagreement_score":0.00443075,"about_ca_system_score_codex":0.00037195574,"about_ca_system_score_gemma":0.00050592603,"threshold_uncertainty_score":0.008809924},"labels":[],"label_agreement":null},{"id":"W2768733040","doi":"10.3934/bdia.2017016","title":"An ontological account of flow-control components in BPMN process models","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Business Process Modeling and Analysis","field":"Business, Management and Accounting","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"TD Bank Group; York University","funders":"","keywords":"Business Process Model and Notation; Petri net; Computer science; XPDL; Programming language; Business process; Process modeling; Business process modeling; Software engineering; Theoretical computer science; Workflow; Database; Work in process; Workflow management system","score_opus":0.14119815110025397,"score_gpt":0.2975243409845658,"score_spread":0.15632618988431185,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2768733040","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0074329684,0.0007737376,0.9409188,0.0021360451,0.00022270427,0.00010642936,0.00017593197,0.00022816325,0.048005275],"genre_scores_gemma":[0.37595955,0.0027919803,0.604193,0.0011604463,0.00072280254,0.0006944479,0.0009138069,0.00024856933,0.013315386],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99685186,0.0009562947,0.00030083745,0.000402618,0.0011586526,0.00032970967],"domain_scores_gemma":[0.9979097,0.0007302405,0.0002811577,0.00050213153,0.0004010071,0.00017572397],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038360895,0.00082629855,0.00040410514,0.0028538338,0.0021866362,0.005396179,0.0016385355,0.002055622,0.0022617113],"category_scores_gemma":[0.004535447,0.000673153,0.0014314327,0.0022583501,0.005695612,0.009466298,0.0030435175,0.0029478762,0.0005594693],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000004285767,0.00001382683,0.000114289134,0.00002424378,0.000003987055,0.00006618896,0.00037880964,0.0031287612,0.00031043522,0.991856,0.00030270914,0.0037964399],"study_design_scores_gemma":[0.0000109842385,0.000018944613,0.00041775688,0.00011212088,0.000036425292,0.00016193242,0.0002506368,0.0574443,0.0009324518,0.8891777,0.0514017,0.000034982717],"about_ca_topic_score_codex":0.0065779053,"about_ca_topic_score_gemma":0.004721859,"teacher_disagreement_score":0.0065779053,"about_ca_system_score_codex":0.0029531224,"about_ca_system_score_gemma":0.0035982481,"threshold_uncertainty_score":0.021426558},"labels":[],"label_agreement":null},{"id":"W2793062537","doi":"10.3934/bdia.2017018","title":"Fuzzy temporal meta-clustering of financial trading volatility patterns","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Stock Market Forecasting Methods","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Saint Mary's University","funders":"","keywords":"Volatility (finance); Cluster analysis; Volatility clustering; Fuzzy clustering; Fuzzy logic; Econometrics; Finance; Computer science; Business; Economics; Artificial intelligence; Autoregressive conditional heteroskedasticity","score_opus":0.4454039739101132,"score_gpt":0.428209438481672,"score_spread":0.017194535428441238,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2793062537","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.34277034,0.0006732005,0.65255344,0.00036604016,0.000066653905,0.0000692355,0.00075880333,0.00031221338,0.0024300136],"genre_scores_gemma":[0.89767945,0.00021819318,0.10029076,0.000033049957,0.000033577042,0.00004136488,0.0007447621,0.000032325854,0.00092651445],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996031,0.000081306134,0.000042076885,0.0000990051,0.000119483906,0.00005507385],"domain_scores_gemma":[0.99875927,0.00043336893,0.00023488922,0.000118810705,0.00037681704,0.00007686777],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010232959,0.00040677487,0.00050341355,0.0021863815,0.0005215997,0.0013129989,0.0010006074,0.0005101147,0.0011491412],"category_scores_gemma":[0.0029921646,0.00020624601,0.00096583547,0.0020309563,0.00031645168,0.0010529747,0.00043471833,0.00048178085,0.0001906962],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00032209724,0.00014698955,0.022604935,0.00016154209,0.00034310805,0.00024158836,0.00044106288,0.7501625,0.006450647,0.030183392,0.0029229799,0.18601914],"study_design_scores_gemma":[0.000002568789,0.000011115773,0.0028261065,0.000010121046,0.000015153841,0.000028869597,0.000047469315,0.9911362,0.0007408241,0.004845785,0.00032422255,0.000011576987],"about_ca_topic_score_codex":0.009193107,"about_ca_topic_score_gemma":0.008569173,"teacher_disagreement_score":0.009193107,"about_ca_system_score_codex":0.0013870959,"about_ca_system_score_gemma":0.000628238,"threshold_uncertainty_score":0.018279195},"labels":[],"label_agreement":null},{"id":"W2896525219","doi":"10.3934/bdia.2018002","title":"An application of PART to the Football Manager data for players clusters analyses to inform club team formation","year":2017,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Cluster analysis; Computer science; Club; Space (punctuation); Cluster (spacecraft); Hierarchical clustering; Football team; Artificial intelligence; Data science; Machine learning; Football; Geography","score_opus":0.2353395288865652,"score_gpt":0.35022244469127956,"score_spread":0.11488291580471435,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2896525219","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.63934386,0.00045852605,0.30410448,0.0028509875,0.00024220788,0.00063615915,0.036365494,0.002472617,0.013525785],"genre_scores_gemma":[0.7624316,0.0001490441,0.20947371,0.00018049928,0.000089859685,0.00029577073,0.024870763,0.00014299435,0.0023657824],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.998796,0.00056191615,0.00006210601,0.00026857716,0.0002427663,0.00006856477],"domain_scores_gemma":[0.996267,0.0019202584,0.00028408214,0.0006758069,0.00069339067,0.00015954862],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0026453792,0.0006591596,0.0004320382,0.002686157,0.00058442564,0.0015658224,0.0010986489,0.0008172651,0.00351603],"category_scores_gemma":[0.0135369245,0.00020758281,0.00075680745,0.0025820718,0.00035112066,0.0010187761,0.0012252972,0.0010214584,0.0010650555],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011943901,0.0009580879,0.29334587,0.0005891971,0.00063710706,0.0007051194,0.0014637776,0.295311,0.011278874,0.017819729,0.033218075,0.34347874],"study_design_scores_gemma":[0.000051299437,0.00023607802,0.08951416,0.00008837708,0.000058892718,0.00019440595,0.0012208826,0.8723875,0.0039875684,0.016528608,0.015656777,0.00007545521],"about_ca_topic_score_codex":0.010953593,"about_ca_topic_score_gemma":0.016298264,"teacher_disagreement_score":0.010953593,"about_ca_system_score_codex":0.0006834356,"about_ca_system_score_gemma":0.0009572971,"threshold_uncertainty_score":0.021779716},"labels":[],"label_agreement":null},{"id":"W2997468375","doi":"10.3934/bdia.2019001","title":"Statistical modeling on human microbiome sequencing data","year":2019,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Oral microbiology and periodontitis research","field":"Dentistry","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Princess Margaret Cancer Centre; University of Toronto; Public Health Ontario","funders":"","keywords":"Microbiome; Human microbiome; Computational biology; Metagenomics; Human Microbiome Project; Biology; Linkage (software); Statistical model; DNA sequencing; Data science; Computer science; Genetics; Artificial intelligence; Gene","score_opus":0.24333426281341072,"score_gpt":0.3871920293009617,"score_spread":0.14385776648755097,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2997468375","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.021521555,0.0016955405,0.9696335,0.0020258303,0.00019848924,0.00012751012,0.0030317635,0.00077914196,0.0009867122],"genre_scores_gemma":[0.613511,0.0070314794,0.34641874,0.001520137,0.0012803834,0.0017115719,0.01667811,0.00051289285,0.011335675],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9920242,0.0050727967,0.00042189617,0.0012340554,0.00088127697,0.00036581847],"domain_scores_gemma":[0.9655147,0.02898693,0.0020119029,0.0013891441,0.0016055993,0.00049171085],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.013208489,0.0014539792,0.0019123782,0.0021910896,0.00068941177,0.0026921327,0.0028518708,0.0019680257,0.0028558536],"category_scores_gemma":[0.033520244,0.00090340246,0.0024787853,0.003243542,0.0014891735,0.002116108,0.0018720308,0.002759997,0.00076836324],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017487176,0.00009139189,0.011764585,0.00029278334,0.0004129868,0.00041536402,0.00019884779,0.849111,0.0004955695,0.08917742,0.005516869,0.042348325],"study_design_scores_gemma":[0.0000108829645,0.000028033788,0.0008707983,0.000019880603,0.00001791736,0.00004546468,0.000023877046,0.9535031,0.00008843271,0.04401881,0.0013579966,0.000014803525],"about_ca_topic_score_codex":0.015020115,"about_ca_topic_score_gemma":0.008962474,"teacher_disagreement_score":0.015020115,"about_ca_system_score_codex":0.0014598235,"about_ca_system_score_gemma":0.0018260368,"threshold_uncertainty_score":0.06985402},"labels":[],"label_agreement":null},{"id":"W3003902258","doi":"10.3934/bdia.2021005","title":"Aggregate loss model with Poisson-Tweedie frequency","year":2021,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Insurance and Financial Risk Management","field":"Economics, Econometrics and Finance","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Wilfrid Laurier University","funders":"","keywords":"Poisson distribution; Percentile; Econometrics; Aggregate (composite); Reinsurance; Compound Poisson distribution; Distribution (mathematics); Zero-inflated model; Sensitivity (control systems); Statistics; Poisson regression; Mathematics; Computer science; Economics; Population; Actuarial science; Engineering","score_opus":0.06624022060634314,"score_gpt":0.2292257438401906,"score_spread":0.16298552323384746,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3003902258","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.26094016,0.0010453054,0.71903837,0.002923713,0.00047127242,0.0003127461,0.0024673655,0.0008155821,0.011985463],"genre_scores_gemma":[0.96588993,0.00057106506,0.017373312,0.00024828786,0.0002047542,0.00023582422,0.0010322478,0.000086919004,0.014357745],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99804807,0.0005466898,0.0001208751,0.0004157966,0.0005435587,0.00032492864],"domain_scores_gemma":[0.988588,0.0068344516,0.0019070209,0.0007551147,0.0014781681,0.00043724844],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0055739037,0.0010695558,0.001258006,0.0021771654,0.0008387156,0.0030219404,0.003927167,0.0030349612,0.0062722825],"category_scores_gemma":[0.015741225,0.0006544386,0.0016339996,0.0022661036,0.002043709,0.0038192535,0.0014855136,0.0032676496,0.0012371877],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000158326,0.00014709172,0.006997586,0.00007277722,0.00006656184,0.0006541563,0.0003150518,0.84896845,0.00080566073,0.1273535,0.004150609,0.01031024],"study_design_scores_gemma":[0.000010825301,0.000021224605,0.0006513045,0.000009673819,0.000010933349,0.00008368292,0.000038531696,0.9856537,0.00011091664,0.012881363,0.00051023334,0.000017599406],"about_ca_topic_score_codex":0.014710544,"about_ca_topic_score_gemma":0.005398565,"teacher_disagreement_score":0.014710544,"about_ca_system_score_codex":0.002479723,"about_ca_system_score_gemma":0.0008549857,"threshold_uncertainty_score":0.029477954},"labels":[],"label_agreement":null},{"id":"W3047104034","doi":"10.3934/bdia.2020001","title":"Modeling portfolio loss by interval distributions","year":2020,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Probabilistic and Robust Engineering Design","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Royal Bank of Canada","funders":"","keywords":"Portfolio; Interval (graph theory); Outcome (game theory); Mathematics; Statistics; Econometrics; Capital allocation line; Regression analysis; Regression; Computer science; Economics; Mathematical economics; Combinatorics","score_opus":0.21734204926902354,"score_gpt":0.3329600770183434,"score_spread":0.11561802774931987,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3047104034","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01654789,0.0006463157,0.9785139,0.00043786198,0.000046271463,0.00005876502,0.00029323722,0.00027013558,0.0031855851],"genre_scores_gemma":[0.82061434,0.0031321466,0.15650697,0.00045868062,0.00030206866,0.00088268175,0.0010436577,0.00031696534,0.016742429],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99544626,0.0023270152,0.00014502509,0.0007462834,0.0009423732,0.00039304706],"domain_scores_gemma":[0.9780563,0.017536458,0.0019733692,0.0010595322,0.00104407,0.00033024556],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011254372,0.001509218,0.0014712261,0.0017873411,0.00033635966,0.0029907096,0.0028725937,0.0019596647,0.0058698137],"category_scores_gemma":[0.038115226,0.0007337756,0.0015491333,0.001957395,0.0017142989,0.004167546,0.0017526721,0.0034122812,0.0011536931],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016797936,0.00011770098,0.0036183372,0.00011580008,0.00011171776,0.00020128729,0.00019574474,0.6939335,0.00053314085,0.26481298,0.002264294,0.033927582],"study_design_scores_gemma":[0.00002492377,0.000071213464,0.0006598167,0.000035427736,0.000028144192,0.000069830676,0.000024953095,0.8426508,0.00019185983,0.15490767,0.0013128413,0.000022480574],"about_ca_topic_score_codex":0.0031543765,"about_ca_topic_score_gemma":0.0013125082,"teacher_disagreement_score":0.011254372,"about_ca_system_score_codex":0.001465563,"about_ca_system_score_gemma":0.0008410677,"threshold_uncertainty_score":0.05951953},"labels":[],"label_agreement":null},{"id":"W4408719487","doi":"10.3934/bdia.2025001","title":"Deep neural networks with application in predicting the spread of avian influenza through disease-informed neural networks","year":2025,"lang":"en","type":"article","venue":"Big Data and Information Analytics","topic":"Influenza Virus Research Studies","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Artificial neural network; Influenza A virus subtype H5N1; Deep neural networks; Disease; Computer science; Artificial intelligence; Virology; Biology; Medicine; Pathology; Virus","score_opus":0.07355436373879289,"score_gpt":0.35757403848675934,"score_spread":0.28401967474796647,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408719487","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6449037,0.0017755487,0.34615004,0.0012141162,0.00016985243,0.000102289014,0.00096404867,0.0013263525,0.0033940147],"genre_scores_gemma":[0.9645582,0.0001698426,0.033632327,0.000109602064,0.000022666422,0.000037640948,0.00047238445,0.000019264298,0.0009781983],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997358,0.00009057909,0.00002358601,0.00006090565,0.000047942132,0.00004117264],"domain_scores_gemma":[0.99877816,0.00079826184,0.0001081698,0.000069315145,0.00020171407,0.000044352382],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016485841,0.0007534849,0.0004306207,0.0006180477,0.00028872065,0.00061254494,0.00080011145,0.0009505144,0.00068724767],"category_scores_gemma":[0.0046739765,0.00044824637,0.00049187517,0.0005716008,0.0003715308,0.00085564016,0.0007835672,0.0013682944,0.00011538261],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000034343466,0.000029522545,0.0022487666,0.00001486171,0.000021740094,0.000024293826,0.000015814414,0.98617446,0.00031994918,0.0006010764,0.00022844202,0.010286649],"study_design_scores_gemma":[0.0000013436571,0.00000606597,0.00018859815,0.0000019336046,0.0000014967123,0.0000022774175,0.0000024269932,0.99911493,0.00014452338,0.00049450644,0.000040307794,0.0000015453973],"about_ca_topic_score_codex":0.017797869,"about_ca_topic_score_gemma":0.01680884,"teacher_disagreement_score":0.017797869,"about_ca_system_score_codex":0.001076692,"about_ca_system_score_gemma":0.0010604701,"threshold_uncertainty_score":0.03538853},"labels":[],"label_agreement":null}]}