{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":29,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":29,"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":"b933c51cc8d0","filters":{"venue":"Journal Of Big Data"}},"results":[{"id":"W3044719873","doi":"10.1186/s40537-020-00329-2","title":"Predictive big data analytics for supply chain demand forecasting: methods, applications, and research opportunities","year":2020,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Forecasting Techniques and Applications","field":"Decision Sciences","cited_by":469,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Supply chain; Computer science; Predictive analytics; Demand forecasting; Supply chain management; Big data; Time series; Data science; Cluster analysis; Analytics; Support vector machine; Data mining; Machine learning; Operations research; Business; Engineering; Marketing","authors":[{"name":"Mahya Seyedan","is_ca":true},{"name":"Fereshteh Mafakheri","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.9012087118397553,"gpt":0.5583508521736629,"spread":0.3428578596660924,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004293393,0.001101192,0.001057042,0.003304975,0.0004260436,0.003043124,0.001339177,0.001169391,0.001449422],"category_scores_gemma":[0.01045201,0.0004934039,0.000843969,0.007935752,0.0008571466,0.003367004,0.001179895,0.002434395,0.0006556623],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008396555,"about_ca_system_score_gemma":0.00133517,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003852888,"about_ca_topic_score_gemma":0.002502764,"domain_scores_codex":[0.9984612,0.0006353348,0.00008908437,0.0002166678,0.0005157207,0.00008198944],"domain_scores_gemma":[0.9874177,0.009620682,0.0005140167,0.0006923856,0.001546843,0.0002083578],"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":[0.000173904,0.0003335325,0.02329741,0.002084658,0.0003219693,0.0001876866,0.000389112,0.133028,0.001375917,0.07835561,0.01882362,0.7416285],"study_design_scores_gemma":[0.00002294815,0.00008381852,0.003850974,0.001008608,0.00007637653,0.0001115959,0.0006328876,0.8365754,0.001460753,0.1316237,0.02449381,0.00005912484],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.04915968,0.1532981,0.756181,0.02117146,0.001211582,0.0002717099,0.001873006,0.001264114,0.01556927],"genre_scores_gemma":[0.6757975,0.1362746,0.1803681,0.00109568,0.002069854,0.0002782065,0.001985257,0.00009819834,0.002032606],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.004293393,"threshold_uncertainty_score":0.02270591,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3081799531","doi":"10.1186/s40537-020-00333-6","title":"Short-term stock market price trend prediction using a comprehensive deep learning system","year":2020,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Stock Market Forecasting Methods","field":"Decision Sciences","cited_by":370,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada; Carleton University","keywords":"Feature engineering; Computer science; Stock market; Stock (firearms); Artificial intelligence; Deep learning; Machine learning; Big data; Personalization; Stock market prediction; Econometrics; Data mining; Economics; World Wide Web","authors":[{"name":"Jingyi Shen","is_ca":true},{"name":"M. Omair Shafiq","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4401202832170943,"gpt":0.4249298837676808,"spread":0.01519039944941347,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005250091,0.0006627903,0.0006520663,0.0009503668,0.0003333674,0.0007135563,0.0007501874,0.0004728862,0.001187733],"category_scores_gemma":[0.001098022,0.0003492942,0.0004185708,0.0007517629,0.0001374057,0.001572269,0.0006821036,0.0006697959,0.0004998189],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006748664,"about_ca_system_score_gemma":0.0009620471,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01053185,"about_ca_topic_score_gemma":0.01439864,"domain_scores_codex":[0.9997537,0.0000161569,0.00002621423,0.00008524873,0.00008470955,0.00003402142],"domain_scores_gemma":[0.9995486,0.00007244611,0.00005032731,0.00006292145,0.000227076,0.00003862271],"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":[0.000519182,0.001174561,0.02520472,0.0001304475,0.0002615315,0.0003642006,0.0001106911,0.2430056,0.04035639,0.00175584,0.009061211,0.6780557],"study_design_scores_gemma":[0.000009157707,0.00006343954,0.002057936,0.000003357219,0.0000199391,0.00001753461,0.000005593452,0.9945269,0.002509523,0.0003820115,0.0003954447,0.000009080706],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.639137,0.0006510237,0.3425198,0.0006039162,0.000180143,0.0001652105,0.001874359,0.01008648,0.004782127],"genre_scores_gemma":[0.9198969,0.0001765588,0.07470538,0.0001732287,0.00005805629,0.00006740591,0.002308057,0.00004010522,0.002574372],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01053185,"threshold_uncertainty_score":0.02094108,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3081491601","doi":"10.1186/s40537-020-00345-2","title":"Prediction of probable backorder scenarios in the supply chain using Distributed Random Forest and Gradient Boosting Machine learning techniques","year":2020,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Forecasting Techniques and Applications","field":"Decision Sciences","cited_by":160,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Toronto Metropolitan University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Gradient boosting; Random forest; Computer science; Boosting (machine learning); Decision tree; Machine learning; CLARITY; Artificial intelligence; Flexibility (engineering); Metric (unit); Supply chain; Tree (set theory); Business process; Process (computing); Data mining; Work in process; Statistics; Mathematics; Operations management; Engineering","authors":[{"name":"Samiul Islam","is_ca":true},{"name":"Saman Hassanzadeh Amin","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3050417918781734,"gpt":0.3606107987158744,"spread":0.055569006837701,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001795316,0.0006654324,0.0006991246,0.001287933,0.0003969006,0.000665325,0.0006816277,0.0007349612,0.0009134432],"category_scores_gemma":[0.003434895,0.0003139371,0.0008361394,0.0007902755,0.000229533,0.0006932856,0.0002967411,0.0007922367,0.0002268767],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008692655,"about_ca_system_score_gemma":0.0007953899,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01227912,"about_ca_topic_score_gemma":0.01058999,"domain_scores_codex":[0.9994088,0.0002099896,0.00003650451,0.0001147763,0.0001317482,0.00009821462],"domain_scores_gemma":[0.9972693,0.001665043,0.0002884415,0.0001260919,0.000494506,0.0001565956],"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":[0.00005510084,0.00005151646,0.004832235,0.00001576319,0.00001356617,0.00006565014,0.00001141988,0.9838436,0.0004437724,0.0003017646,0.0001676187,0.01019802],"study_design_scores_gemma":[0.000001037651,0.000009845351,0.0005038902,0.000001508596,0.000001404369,0.000003918721,0.000003760871,0.9990792,0.000133389,0.0002428724,0.00001748545,0.000001636874],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6890613,0.0003196748,0.3076929,0.0003011669,0.00004524544,0.00008043167,0.000472037,0.0007054317,0.001321801],"genre_scores_gemma":[0.9639474,0.0000525315,0.03519288,0.00001804235,0.000008742767,0.00002265421,0.0003814205,0.00001059476,0.0003656059],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01227912,"threshold_uncertainty_score":0.02441531,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4210862833","doi":"10.1186/s40537-021-00468-0","title":"Big data quality framework: a holistic approach to continuous quality management","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":141,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Big data; Computer science; Data quality; Profiling (computer programming); Data science; Data mining; Quality (philosophy); Data management; Exploratory data analysis; Engineering; Operations management","authors":[{"name":"Ikbal Taleb","is_ca":false},{"name":"Mohamed Adel Serhani","is_ca":false},{"name":"Chafik Bouhaddioui","is_ca":false},{"name":"Rachida Dssouli","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7948772270531455,"gpt":0.5341265486844159,"spread":0.2607506783687296,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02319967,0.001635723,0.001460026,0.008474006,0.001729945,0.01249012,0.005137668,0.002282169,0.001491374],"category_scores_gemma":[0.01704438,0.0009524018,0.002635168,0.006920575,0.004829486,0.009860926,0.007141108,0.004336807,0.0004441428],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004613797,"about_ca_system_score_gemma":0.008362233,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008350067,"about_ca_topic_score_gemma":0.004430864,"domain_scores_codex":[0.9801129,0.007178219,0.002191905,0.002336131,0.007177746,0.001003089],"domain_scores_gemma":[0.9840884,0.004016307,0.002351393,0.002461807,0.005660739,0.001421214],"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":[0.0001093133,0.0002621348,0.007876138,0.001075091,0.0002837821,0.0005295757,0.001807496,0.08542763,0.002813424,0.7666662,0.008924868,0.1242244],"study_design_scores_gemma":[0.00004817708,0.0002370221,0.003071013,0.001147669,0.000230517,0.0004890858,0.001600674,0.5202304,0.003794894,0.4036711,0.06530437,0.000175027],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003223868,0.001045924,0.9877215,0.002229395,0.0001161466,0.0003257338,0.0002040851,0.0007506341,0.004382772],"genre_scores_gemma":[0.1608619,0.001365987,0.834285,0.0005096621,0.0002860672,0.000522727,0.0007481945,0.0001623917,0.001258121],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02319967,"threshold_uncertainty_score":0.1226931,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2132987460","doi":"10.1186/s40537-014-0011-y","title":"Contextual anomaly detection framework for big sensor data","year":2015,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":141,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Anomaly detection; Context (archaeology); Toolbox; Data mining; Detector; Cluster analysis; Task (project management); False positive paradox; Anomaly (physics); Volume (thermodynamics); Big data; Process (computing); Content (measure theory); Artificial intelligence; Machine learning; Pattern recognition (psychology)","authors":[{"name":"Miriam A. M. Capretz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2992592294446529,"gpt":0.3639286034466015,"spread":0.06466937400194861,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002477812,0.001124858,0.001301575,0.003100538,0.0007424194,0.001932723,0.00296418,0.001049813,0.001431416],"category_scores_gemma":[0.006810115,0.0004322017,0.001441314,0.00236522,0.0007556896,0.002108391,0.002232782,0.001902372,0.0006145571],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001163522,"about_ca_system_score_gemma":0.001933544,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008479136,"about_ca_topic_score_gemma":0.01139873,"domain_scores_codex":[0.997947,0.0003749265,0.0001433225,0.0004734047,0.0009136513,0.0001476699],"domain_scores_gemma":[0.9978415,0.0006730731,0.0002691049,0.0003895077,0.0007005615,0.0001263293],"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":[0.0004163756,0.0004123454,0.01142425,0.0006550093,0.0005425002,0.0007695776,0.0005458225,0.3987662,0.01472475,0.1041396,0.02005054,0.447553],"study_design_scores_gemma":[0.00001293809,0.00004118557,0.000972289,0.00001819475,0.00003635816,0.0001413239,0.00006426867,0.9535026,0.002801675,0.03465028,0.007732538,0.00002625733],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002657001,0.0002966749,0.9932903,0.0001405434,0.00004852668,0.00006793653,0.0002854776,0.002754059,0.0004593838],"genre_scores_gemma":[0.196153,0.0005161325,0.7993755,0.0002238052,0.0002211233,0.0002886089,0.001624081,0.0003283332,0.001269414],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008479136,"threshold_uncertainty_score":0.01685959,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4391520039","doi":"10.1186/s40537-023-00842-0","title":"Survey of transformers and towards ensemble learning using transformers for natural language processing","year":2024,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Topic Modeling","field":"Computer Science","cited_by":112,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada; Carleton University","keywords":"Computer science; Transformer; Automatic summarization; Artificial intelligence; Language model; Question answering; Natural language processing; Classifier (UML); Natural language; Machine learning; Sentiment analysis; Natural language understanding; Ensemble forecasting","authors":[{"name":"Hongzhi Zhang","is_ca":true},{"name":"M. Omair Shafiq","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1431906889705772,"gpt":0.3509968199453188,"spread":0.2078061309747416,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002867182,0.001622496,0.00129751,0.002399316,0.0004586726,0.001871813,0.002085504,0.000878491,0.002567156],"category_scores_gemma":[0.00669644,0.000673644,0.001742522,0.003141545,0.0007256471,0.00590084,0.001983661,0.00233473,0.001412829],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001185371,"about_ca_system_score_gemma":0.001501766,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005278216,"about_ca_topic_score_gemma":0.00388808,"domain_scores_codex":[0.9986107,0.0004681699,0.0001138897,0.000364155,0.0003644769,0.00007852579],"domain_scores_gemma":[0.9971454,0.001653009,0.0001003108,0.0004588784,0.0005463647,0.00009612323],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001918308,0.0002106576,0.004248561,0.0007025637,0.0003189522,0.0001012272,0.0002205667,0.1066788,0.002668582,0.05633303,0.01414452,0.8141807],"study_design_scores_gemma":[0.00001504081,0.0002163902,0.0008394908,0.0001134464,0.0001161121,0.0001761005,0.0000831174,0.9062521,0.002761169,0.0713103,0.01807844,0.0000382828],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.01197241,0.02539994,0.9537063,0.001263393,0.0002997178,0.0001004238,0.0002966622,0.002279059,0.004682138],"genre_scores_gemma":[0.5181679,0.07933551,0.3833855,0.001767161,0.001341791,0.0004660132,0.0036896,0.0007853341,0.01106116],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.005278216,"threshold_uncertainty_score":0.01516324,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4360827072","doi":"10.1186/s40537-023-00711-w","title":"Deep learning based deep-sea automatic image enhancement and animal species classification","year":2023,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Ichthyology and Marine Biology","field":"Environmental Science","cited_by":41,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Centro para el Desarrollo Tecnológico Industrial; Ministerio de Ciencia, Innovación y Universidades","keywords":"Computer science; Artificial intelligence; Residual; Deep learning; Pipeline (software); Pattern recognition (psychology); Data set; Set (abstract data type); Convolutional neural network; Algorithm","authors":[{"name":"Vanesa López-Vázquez","is_ca":false},{"name":"José Manuel López-Guede","is_ca":false},{"name":"Damianos Chatzievangelou","is_ca":false},{"name":"Jacopo Aguzzi","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06251843455730097,"gpt":0.2822145933443102,"spread":0.2196961587870092,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005757501,0.0007864629,0.0004663749,0.0008245634,0.000160302,0.0004323575,0.0007230844,0.0005596656,0.001783625],"category_scores_gemma":[0.0006856932,0.0002346821,0.0006107885,0.0004609766,0.0002589651,0.0005360699,0.0005854816,0.0006754134,0.0008670218],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004177731,"about_ca_system_score_gemma":0.0004397337,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004357764,"about_ca_topic_score_gemma":0.005484404,"domain_scores_codex":[0.9997795,0.00002711907,0.0000115297,0.00005280972,0.00006568705,0.00006335076],"domain_scores_gemma":[0.9997039,0.0000619831,0.00004168268,0.00004493658,0.0001285029,0.00001892069],"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":[0.0004360798,0.0003441013,0.005137005,0.000174528,0.0001462106,0.0001853611,0.00008441008,0.1238058,0.1483732,0.001234852,0.00483715,0.7152413],"study_design_scores_gemma":[0.000008311487,0.0001027265,0.0027513,0.0000132254,0.00002382595,0.00004934013,0.00001748002,0.9462594,0.04899123,0.0004649015,0.001305169,0.00001309535],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2847471,0.0009340579,0.6994057,0.0003462416,0.0001472674,0.000160814,0.0007805965,0.009109602,0.004368703],"genre_scores_gemma":[0.7578669,0.0003249049,0.2320378,0.0002266038,0.00003683772,0.00008429012,0.001472242,0.0001643687,0.007786071],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004357764,"threshold_uncertainty_score":0.008664787,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4384821779","doi":"10.1186/s40537-023-00796-3","title":"Detecting bots in social-networks using node and structural embeddings","year":2023,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Broadcom (Canada); Toronto Metropolitan University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Metadata; Embedding; Class (philosophy); Node (physics); Machine learning; Social network (sociolinguistics); Feature (linguistics); Artificial intelligence; Anonymity; Focus (optics); Information retrieval; Data mining; Social media; World Wide Web","authors":[{"name":"Ashkan Dehghan","is_ca":true},{"name":"Kinga Siuta","is_ca":true},{"name":"Agata Skorupka","is_ca":true},{"name":"Akshat Dubey","is_ca":true},{"name":"Andrei Betlen","is_ca":true},{"name":"David Miller","is_ca":true},{"name":"Wei Xu","is_ca":true},{"name":"Bogumił Kamiński","is_ca":false},{"name":"Paweł Prałat","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1361918284243414,"gpt":0.3333503352020518,"spread":0.1971585067777104,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007925448,0.0005837838,0.0004677115,0.003542539,0.0003759103,0.001020652,0.0003876438,0.0008557437,0.001089002],"category_scores_gemma":[0.006232741,0.0002348969,0.0004294149,0.001763674,0.0006800059,0.002594994,0.0009898961,0.0007611448,0.000582143],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004725935,"about_ca_system_score_gemma":0.000282712,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002169687,"about_ca_topic_score_gemma":0.003233689,"domain_scores_codex":[0.9993796,0.0002202754,0.00004471582,0.0001468865,0.0001293514,0.00007909301],"domain_scores_gemma":[0.9939305,0.003372623,0.001426621,0.0005125747,0.0004969792,0.0002608441],"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":[0.0007952332,0.0006794914,0.4586076,0.0005744977,0.0004004407,0.0008239067,0.001493944,0.2396473,0.01605803,0.02059715,0.006581163,0.2537412],"study_design_scores_gemma":[0.000006798169,0.00009175419,0.02670928,0.00004914959,0.00002681205,0.0002270107,0.0004573604,0.9550988,0.002250335,0.0138626,0.00119614,0.00002394334],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8573999,0.0005491958,0.1366076,0.0005825138,0.00004917138,0.00009873223,0.001439767,0.0006781028,0.002594935],"genre_scores_gemma":[0.9850549,0.0001059736,0.013602,0.00002141213,0.00002490786,0.00002134389,0.0006790835,0.00001745733,0.0004730412],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003542539,"threshold_uncertainty_score":0.004314065,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4324378562","doi":"10.1186/s40537-023-00710-x","title":"A semi-supervised short text sentiment classification method based on improved Bert model from unlabelled data","year":2023,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Bottleneck; Sentiment analysis; Artificial intelligence; Semi-supervised learning; Machine learning; Big data; Supervised learning; Language model; Function (biology); Natural language processing; Data mining; Artificial neural network","authors":[{"name":"Haochen Zou","is_ca":false},{"name":"Zitao Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2707046506402733,"gpt":0.3750577178151738,"spread":0.1043530671749005,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009634281,0.001057154,0.0009585054,0.001172964,0.0006058888,0.0008097489,0.001787452,0.0009995655,0.002227114],"category_scores_gemma":[0.001627171,0.0003887465,0.0009008281,0.0007097026,0.0003929238,0.001583818,0.0007188633,0.001035018,0.001178084],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000794247,"about_ca_system_score_gemma":0.0008863439,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00639329,"about_ca_topic_score_gemma":0.007446339,"domain_scores_codex":[0.9994112,0.00009987212,0.00004396329,0.0001726903,0.0001902004,0.00008214654],"domain_scores_gemma":[0.9992041,0.0001792415,0.00006884312,0.00007104863,0.0004241559,0.00005256534],"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":[0.0008885717,0.0005879161,0.008885298,0.0002020723,0.0001521733,0.0003386301,0.000261733,0.2325515,0.03202963,0.003731552,0.01346283,0.706908],"study_design_scores_gemma":[0.000005994061,0.00002642294,0.0003550702,0.000003307843,0.000007490577,0.0000140565,0.000009585263,0.9974112,0.001453102,0.0004174055,0.0002907176,0.000005699204],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1452592,0.0006761482,0.8440723,0.0006695605,0.0003192094,0.0002584615,0.0005136308,0.003615922,0.004615513],"genre_scores_gemma":[0.8554309,0.0003595477,0.1295015,0.0004128112,0.000295052,0.0003256378,0.00248665,0.0001778676,0.01101],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00639329,"threshold_uncertainty_score":0.01271212,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3155713909","doi":"10.1186/s40537-021-00455-5","title":"Domain randomization for neural network classification","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":26,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Computer science; Classifier (UML); Convolutional neural network; Artificial intelligence; Artificial neural network; Task (project management); Domain (mathematical analysis); Machine learning; Pattern recognition (psychology); Randomization; Transfer of learning; Contextual image classification; Data mining; Image (mathematics); Clinical trial; Mathematics","authors":[{"name":"Svetozar Zarko Valtchev","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1690613487970136,"gpt":0.3181836121059016,"spread":0.149122263308888,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002752468,0.0007010297,0.0008125714,0.0005964385,0.000454568,0.0008933873,0.001345253,0.001144408,0.00219422],"category_scores_gemma":[0.01188648,0.0003977256,0.0007602952,0.000694101,0.001105957,0.001804967,0.001343205,0.002444146,0.0009318201],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001010504,"about_ca_system_score_gemma":0.0008148433,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001582684,"about_ca_topic_score_gemma":0.001484896,"domain_scores_codex":[0.9988701,0.0005213528,0.00006033463,0.0003126499,0.0001684736,0.00006700955],"domain_scores_gemma":[0.9967043,0.001784202,0.0002413238,0.0009003299,0.0002763278,0.00009349998],"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":[0.0004214634,0.0002286327,0.003496773,0.0002761554,0.0001362664,0.0001465988,0.00009652542,0.7257523,0.009950066,0.0435123,0.009505554,0.2064774],"study_design_scores_gemma":[0.00001272059,0.00004114962,0.0003314008,0.00001495884,0.000005716495,0.00003651288,0.00001170917,0.9664794,0.002371933,0.02908611,0.001597691,0.00001060522],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04178901,0.001329385,0.9510252,0.0006740045,0.0001691504,0.0001305583,0.0005122534,0.002135514,0.002234954],"genre_scores_gemma":[0.659229,0.0008071466,0.3328438,0.0006010117,0.0001943211,0.0006035574,0.002115331,0.0003473205,0.003258478],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002752468,"threshold_uncertainty_score":0.01455659,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3022470673","doi":"10.1186/s40537-020-00302-z","title":"Leveraging machine learning and big data for optimizing medication prescriptions in complex diseases: a case study in diabetes management","year":2020,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Diabetes, Cardiovascular Risks, and Lipoproteins","field":"Medicine","cited_by":24,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University; Toronto Metropolitan University","funders":"","keywords":"Computer science; Bayesian network; Big data; Machine learning; Medical prescription; Variable (mathematics); Artificial intelligence; Bayesian probability; Data mining; Medicine; Mathematics","authors":[{"name":"Mahsa Madani Hosseini","is_ca":true},{"name":"Manaf Zargoush","is_ca":true},{"name":"Farrokh Alemi","is_ca":false},{"name":"Raya Kheirbek","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2050337070575335,"gpt":0.3363276992871178,"spread":0.1312939922295843,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004402338,0.0006552637,0.0008941636,0.0009856279,0.0006980873,0.001325374,0.0008824201,0.001708602,0.0009337305],"category_scores_gemma":[0.009459522,0.0004535978,0.0008328418,0.001260987,0.0007586219,0.00108901,0.001110025,0.001354287,0.00008284554],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001257751,"about_ca_system_score_gemma":0.002011255,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009634621,"about_ca_topic_score_gemma":0.009759998,"domain_scores_codex":[0.99845,0.001012413,0.00008043316,0.0001681958,0.0001840257,0.0001048186],"domain_scores_gemma":[0.9915836,0.007183243,0.000415313,0.000246967,0.0003038203,0.0002671735],"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":[0.0002762669,0.0005834468,0.02548294,0.0001182419,0.0001430451,0.001003055,0.0001954859,0.88504,0.0008458607,0.007900987,0.001213277,0.0771973],"study_design_scores_gemma":[0.00005246572,0.00007797935,0.001659305,0.00001269903,0.00002287564,0.0001056203,0.00007410264,0.9926117,0.0005695024,0.00422907,0.0005716234,0.00001303593],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6112227,0.001413953,0.3774711,0.005253794,0.00007716327,0.0002963748,0.0003068743,0.0002751867,0.003682916],"genre_scores_gemma":[0.8495392,0.0002634925,0.1492266,0.0001732251,0.00003731943,0.00009379313,0.0001120686,0.00001929824,0.0005349681],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009634621,"threshold_uncertainty_score":0.02328205,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1623534659","doi":"10.1186/s40537-015-0021-4","title":"Meta-MapReduce for scalable data mining","year":2015,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa; Dalhousie University","funders":"","keywords":"Computer science; Scalability; Big data; AdaBoost; Machine learning; Node (physics); Programming paradigm; Cloud computing; Artificial intelligence; Data mining; Database; Programming language; Operating system; Support vector machine","authors":[{"name":"Xuan Liu","is_ca":true},{"name":"Xiaoguang Wang","is_ca":true},{"name":"Stan Matwin","is_ca":true},{"name":"Nathalie Japkowicz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6436468009513665,"gpt":0.3948568341215879,"spread":0.2487899668297786,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002677185,0.001198731,0.001137231,0.001518531,0.001166552,0.001913957,0.003685373,0.0006839767,0.003174375],"category_scores_gemma":[0.003891065,0.0008916737,0.002562968,0.002596216,0.0006093958,0.002530038,0.002819138,0.001828754,0.002094832],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001034882,"about_ca_system_score_gemma":0.002577693,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004964501,"about_ca_topic_score_gemma":0.006088729,"domain_scores_codex":[0.9979462,0.0004733485,0.0001910354,0.0004267881,0.0007815899,0.0001810366],"domain_scores_gemma":[0.998181,0.0004029113,0.00009125242,0.0007890814,0.0003950501,0.0001407623],"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":[0.0009826651,0.0006255988,0.005034094,0.001414922,0.001016754,0.0007299302,0.0005439346,0.1673066,0.02273887,0.07372227,0.1630884,0.5627959],"study_design_scores_gemma":[0.0001964541,0.0001296876,0.001744269,0.00007122055,0.00008551464,0.0004104926,0.0002344656,0.7892327,0.01568955,0.1240694,0.06805522,0.00008101507],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01014276,0.001545043,0.9531105,0.001190411,0.0003730585,0.0005013425,0.002428672,0.02400906,0.006699136],"genre_scores_gemma":[0.1912597,0.001127204,0.7933522,0.0005010669,0.0002302644,0.0008690856,0.006993419,0.001416613,0.004250353],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004964501,"threshold_uncertainty_score":0.01415849,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4302614395","doi":"10.1186/s40537-022-00608-0","title":"Adaptive multiple imputations of missing values using the class center","year":2022,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba","funders":"Thailand Research Fund; Natural Sciences and Engineering Research Council of Canada; Khon Kaen University","keywords":"Imputation (statistics); Missing data; Categorical variable; Computer science; Data mining; Artificial intelligence; Machine learning","authors":[{"name":"Kritbodin Phiwhorm","is_ca":false},{"name":"Charnnarong Saikaew","is_ca":false},{"name":"Carson K. Leung","is_ca":true},{"name":"Pattarawit Polpinit","is_ca":false},{"name":"Kanda Runapongsa Saikaew","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1952227141855969,"gpt":0.3341741212312561,"spread":0.1389514070456592,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01139377,0.001042766,0.002354206,0.002461229,0.001441112,0.001537889,0.004029055,0.001457461,0.002179535],"category_scores_gemma":[0.0263627,0.0005064394,0.002273383,0.003063417,0.0009679542,0.001962576,0.00191393,0.002583719,0.0007885259],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009786697,"about_ca_system_score_gemma":0.002598393,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00434794,"about_ca_topic_score_gemma":0.003290971,"domain_scores_codex":[0.9922104,0.003431974,0.0004648808,0.001996411,0.001531445,0.0003648988],"domain_scores_gemma":[0.9765385,0.01085183,0.002408958,0.004986336,0.004696288,0.0005181684],"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":[0.001783697,0.0006921304,0.02849117,0.0005095007,0.0009569022,0.0004870526,0.0007263155,0.1839453,0.006714864,0.02013231,0.01435702,0.7412037],"study_design_scores_gemma":[0.00008864654,0.0001586727,0.003353939,0.00006567504,0.0001467606,0.0002173655,0.0001406428,0.9649383,0.009794086,0.01596033,0.005056678,0.00007893913],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01718369,0.0004366473,0.9798456,0.0001897655,0.0001235544,0.0001095359,0.0002038904,0.001412452,0.0004948528],"genre_scores_gemma":[0.3231455,0.0003264682,0.6731293,0.0002462292,0.0001992444,0.00031363,0.001254781,0.0002559941,0.001128894],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01139377,"threshold_uncertainty_score":0.06025672,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2947950257","doi":"10.1186/s40537-019-0203-6","title":"Detecting taxi movements using Random Swap clustering and sequential pattern mining","year":2019,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University","funders":"Carleton University","keywords":"Cluster analysis; Computer science; Swap (finance); Silhouette; Data mining; Hierarchical clustering; k-means clustering; Artificial intelligence","authors":[{"name":"Rami Ibrahim","is_ca":true},{"name":"M. Omair Shafiq","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1150805794770239,"gpt":0.2957330234356438,"spread":0.18065244395862,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006399604,0.0008251789,0.000699029,0.004483629,0.0006652285,0.0007451881,0.0009909585,0.0005553063,0.0008533103],"category_scores_gemma":[0.002250618,0.0003290438,0.0008562754,0.003617974,0.0003133095,0.0008793217,0.0007228471,0.0003913816,0.0005948793],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006673876,"about_ca_system_score_gemma":0.0010116,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02104358,"about_ca_topic_score_gemma":0.02706612,"domain_scores_codex":[0.9991874,0.0001162518,0.00007500951,0.0002781946,0.0002478204,0.00009522343],"domain_scores_gemma":[0.9989956,0.0002081394,0.0001615132,0.0001658477,0.0004052686,0.00006364952],"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":[0.000769585,0.0005236405,0.07567625,0.0004232017,0.0004913834,0.0007465979,0.0006581994,0.2105328,0.02093085,0.005384438,0.01393229,0.6699308],"study_design_scores_gemma":[0.0000206714,0.00007303961,0.01248203,0.0000155663,0.000045679,0.0002270585,0.0002494644,0.9746296,0.006573625,0.002784106,0.00286965,0.00002945464],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2671353,0.0003625706,0.7183812,0.0003543238,0.0001000332,0.000425864,0.003062599,0.006759562,0.003418567],"genre_scores_gemma":[0.5952876,0.0001865641,0.3981985,0.0000635523,0.00003279482,0.0002124252,0.004375211,0.0001460374,0.001497307],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02104358,"threshold_uncertainty_score":0.04184216,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3181834696","doi":"10.1186/s40537-021-00489-9","title":"Examining the impact of cross-domain learning on crime prediction","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Crime Patterns and Interventions","field":"Social Sciences","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Memorial University of Newfoundland; Dalhousie University","funders":"Natural Sciences and Engineering Research Council of Canada; Memorial University of Newfoundland; Dalhousie University","keywords":"Computer science; Generalization; Perspective (graphical); Domain (mathematical analysis); Transfer of learning; Data science; Domain knowledge; Task (project management); Machine learning; Baseline (sea); Crime analysis; Demographics; Artificial intelligence; Subject-matter expert; Data mining; Expert system","authors":[{"name":"Fateha Khanam Bappee","is_ca":true},{"name":"Amílcar Soares","is_ca":true},{"name":"Lucas May Petry","is_ca":false},{"name":"Stan Matwin","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.34207003076943,"gpt":0.4718022365358112,"spread":0.1297322057663812,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01716483,0.001660105,0.001051569,0.002112144,0.0009234793,0.001884977,0.001331059,0.00164151,0.001176273],"category_scores_gemma":[0.02761726,0.0003875207,0.001150158,0.002058299,0.0008646511,0.003654915,0.002874128,0.002623158,0.0004089571],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001558037,"about_ca_system_score_gemma":0.001740248,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02413159,"about_ca_topic_score_gemma":0.01609326,"domain_scores_codex":[0.99472,0.003126692,0.0002394584,0.001111704,0.000457438,0.000344753],"domain_scores_gemma":[0.9734823,0.01941883,0.001000131,0.002954289,0.002446898,0.0006975499],"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":[0.001293253,0.002165729,0.1753616,0.0002156729,0.001047345,0.0001854568,0.000312764,0.6326326,0.001229436,0.001651273,0.003709082,0.1801959],"study_design_scores_gemma":[0.00002320864,0.000257361,0.01104173,0.00002803343,0.00008562422,0.00003643975,0.000225304,0.9848478,0.001224813,0.001720285,0.000488381,0.00002103148],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.91434,0.002274659,0.07557656,0.001300051,0.0002166569,0.0001935645,0.0008694561,0.0007940718,0.004434934],"genre_scores_gemma":[0.9796139,0.0002351379,0.0179853,0.0001394947,0.00004592885,0.00004400823,0.001372133,0.00002950949,0.0005346472],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02413159,"threshold_uncertainty_score":0.09077734,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2772611570","doi":"10.1186/s40537-017-0106-3","title":"A computing platform for pairs-trading online implementation via a blended Kalman-HMM filtering approach","year":2017,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Complex Systems and Time Series Analysis","field":"Economics, Econometrics and Finance","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University; TD Bank Group","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Trading strategy; Kalman filter; Hidden Markov model; Pairs trade; Algorithmic trading; Analytics; High-frequency trading; Big data; Herding; Algorithm; Machine learning; Data mining; Artificial intelligence; Econometrics; Finance; Economics","authors":[{"name":"Anton Tenyakov","is_ca":true},{"name":"Rogemar Mamon","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.270147184667116,"gpt":0.3335862340197416,"spread":0.06343904935262562,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000942642,0.0006066204,0.0007535424,0.0006033815,0.0007132292,0.001501841,0.001983994,0.001103551,0.008423064],"category_scores_gemma":[0.002843648,0.0005453858,0.0007259695,0.0004812489,0.0006291236,0.002213242,0.001995523,0.00134735,0.003468531],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005930487,"about_ca_system_score_gemma":0.0009767319,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003510155,"about_ca_topic_score_gemma":0.003406016,"domain_scores_codex":[0.9991719,0.0001399912,0.00007002917,0.0002516262,0.0002738458,0.00009269601],"domain_scores_gemma":[0.9990921,0.0002659551,0.00005247751,0.0003393728,0.0001752035,0.00007486293],"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":[0.0007495907,0.0004026395,0.002631149,0.0001748984,0.0001494968,0.000431923,0.000536577,0.2705411,0.03800393,0.06747063,0.006377433,0.6125306],"study_design_scores_gemma":[0.00002428718,0.00006587631,0.000218549,0.00001417156,0.00001989284,0.000051808,0.00003594633,0.9702901,0.008121744,0.01471507,0.006418946,0.00002360117],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002986575,0.00003624115,0.9925963,0.000059097,0.00003078132,0.00003335661,0.00002701079,0.002827786,0.001402892],"genre_scores_gemma":[0.2709111,0.0001193965,0.7230574,0.0001353807,0.00005853709,0.0002510628,0.0002492825,0.0003563654,0.004861479],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008423064,"threshold_uncertainty_score":0.02817792,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2952066860","doi":"10.1186/s40537-019-0218-z","title":"STVG: an evolutionary graph framework for analyzing fast-evolving networks","year":2019,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Transportation Planning and Optimization","field":"Social Sciences","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada; Tertiary Education Trust Fund; Cisco Systems","keywords":"Computer science; Theoretical computer science; Graph; Snapshot (computer storage); Graph database; Database","authors":[{"name":"Iyke Maduako","is_ca":true},{"name":"Mónica Wachowicz","is_ca":true},{"name":"Trevor Hanson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0752406044275541,"gpt":0.3422732496091298,"spread":0.2670326451815757,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008572858,0.001223787,0.0009107899,0.00381415,0.0006578251,0.001287572,0.001595267,0.0009446727,0.002315975],"category_scores_gemma":[0.003653735,0.0005072058,0.00143278,0.003054317,0.0007784393,0.001679373,0.001152629,0.001221491,0.0004619659],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009891018,"about_ca_system_score_gemma":0.0009505226,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01201134,"about_ca_topic_score_gemma":0.01371388,"domain_scores_codex":[0.9996412,0.0001189798,0.00001978784,0.0001059499,0.00008132451,0.00003279355],"domain_scores_gemma":[0.9988781,0.0005934698,0.0001630259,0.0001424688,0.0001570074,0.0000659085],"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":[0.00003854928,0.0000462692,0.003546858,0.0001582817,0.0001615178,0.0002022664,0.0001925989,0.8697246,0.003309951,0.05803043,0.003950873,0.06063781],"study_design_scores_gemma":[0.000003800808,0.0000124158,0.0004377016,0.00001044543,0.00001246308,0.00004183779,0.00003653928,0.9650011,0.0003229225,0.03165723,0.00245341,0.00001018301],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006652376,0.0002024908,0.9908421,0.0001001486,0.00002689177,0.00003599248,0.0005400156,0.0009688389,0.0006311826],"genre_scores_gemma":[0.2696203,0.001047219,0.7212587,0.0001757198,0.00009711345,0.0003648204,0.004250867,0.0007641038,0.002421164],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01201134,"threshold_uncertainty_score":0.02388287,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4415383695","doi":"10.1186/s40537-025-01280-w","title":"Digital twins in healthcare: a review of AI-powered practical applications across health domains","year":2025,"lang":"en","type":"review","venue":"Journal Of Big Data","topic":"Digital Transformation in Industry","field":"Engineering","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"","keywords":"Health care; Digital health; Mental healthcare; Selection (genetic algorithm); Mental health; Virtual patient","authors":[{"name":"Ziad Elgammal","is_ca":false},{"name":"M. Taleb Albrijawi","is_ca":false},{"name":"Reda Alhajj","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1670042130554316,"gpt":0.4465265295323191,"spread":0.2795223164768875,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002737951,0.000793467,0.001265242,0.004526147,0.0004928156,0.002027357,0.0009180719,0.001477937,0.005557891],"category_scores_gemma":[0.006566551,0.000447482,0.001066755,0.004956308,0.0009331965,0.00229969,0.001547447,0.001570174,0.001503363],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0010069,"about_ca_system_score_gemma":0.002749064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002018808,"about_ca_topic_score_gemma":0.002746893,"domain_scores_codex":[0.9988011,0.0003949395,0.0002177632,0.000145522,0.0003784394,0.00006227227],"domain_scores_gemma":[0.9932078,0.005647413,0.0002915711,0.0001293439,0.0006127497,0.0001111814],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0000428906,0.00006093594,0.0003043527,0.04948247,0.0001364073,0.0001235109,0.0002889192,0.0005135898,0.0004869775,0.01339349,0.01151361,0.9236529],"study_design_scores_gemma":[0.00001610583,0.0001309672,0.001160547,0.05036,0.0002728851,0.0007807541,0.0003710966,0.0002569988,0.0005646045,0.006371987,0.9396661,0.00004803089],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0001695591,0.996842,0.0004272408,0.0005413766,0.0001317752,0.00001583977,0.00002343611,0.000009627309,0.001839125],"genre_scores_gemma":[0.001492041,0.9972726,0.0006030484,0.0002642412,0.00009674066,0.00001731889,0.00002533696,0.000004023675,0.0002245616],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.005557891,"threshold_uncertainty_score":0.01859301,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4390759017","doi":"10.1186/s40537-023-00866-6","title":"RPf-GCNs: reciprocal perspective driven fused GCNs for rumor detection on social media","year":2024,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Misinformation and Its Impacts","field":"Social Sciences","cited_by":10,"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":"Institute for Information and Communications Technology Promotion; Ministry of Science and ICT, South Korea; National Research Foundation of Korea; National Research Foundation","keywords":"Reciprocal; Computer science; Rumor; Perspective (graphical); Graph; Social media; Convolutional neural network; Profiling (computer programming); Artificial intelligence; Machine learning; Data science; World Wide Web; Theoretical computer science","authors":[{"name":"Zafran Khan","is_ca":false},{"name":"Jeonghwan Gwak","is_ca":false},{"name":"Naima Iltaf","is_ca":false},{"name":"Witold Pedrycz","is_ca":true},{"name":"Moongu Jeon","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.29748346181366,"gpt":0.4248955531956444,"spread":0.1274120913819843,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005727379,0.001026083,0.0005946055,0.001106914,0.0004427797,0.0006699561,0.001341096,0.001249792,0.001227934],"category_scores_gemma":[0.002639962,0.0002663661,0.0006000189,0.0007246468,0.0004354331,0.001415976,0.0008380001,0.001344345,0.0004368001],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009326373,"about_ca_system_score_gemma":0.000841594,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02155726,"about_ca_topic_score_gemma":0.02580823,"domain_scores_codex":[0.9996839,0.00006070192,0.00001419135,0.00009199124,0.00008406548,0.00006525881],"domain_scores_gemma":[0.9993197,0.000238922,0.00008102412,0.00008690936,0.0002283324,0.0000450885],"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":[0.0005378124,0.0003084861,0.009306054,0.000196016,0.0002410798,0.0003827397,0.0002525705,0.3943212,0.0152308,0.006020453,0.009140045,0.5640627],"study_design_scores_gemma":[0.000004454112,0.00003112365,0.0006935883,0.00001216337,0.00001869509,0.00003663715,0.00002886381,0.9941816,0.002256759,0.002186219,0.0005419209,0.000008032211],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2535589,0.003696692,0.7241119,0.001201613,0.0003875349,0.0002194765,0.001459486,0.007455118,0.007909257],"genre_scores_gemma":[0.8982186,0.0005401096,0.0953486,0.0003258091,0.0001090693,0.00005924757,0.00131441,0.00009473081,0.00398933],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02155726,"threshold_uncertainty_score":0.04286361,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4393112045","doi":"10.1186/s40537-024-00898-6","title":"Multi-sample $$\\zeta $$-mixup: richer, more realistic synthetic samples from a p-series interpolant","year":2024,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada; Compute Canada; Simon Fraser University; Nvidia","keywords":"Algorithm; Computer science; Machine learning; Artificial intelligence; Series (stratigraphy)","authors":[{"name":"Kumar Abhishek","is_ca":true},{"name":"Colin J. Brown","is_ca":false},{"name":"Ghassan Hamarneh","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1817104695287353,"gpt":0.3345108908799068,"spread":0.1528004213511714,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00292916,0.001264854,0.0008412404,0.0006604522,0.0005567783,0.001336505,0.001431934,0.001649029,0.005790364],"category_scores_gemma":[0.009177419,0.0004493377,0.0009997445,0.0005413959,0.001384599,0.001738583,0.002421366,0.002755652,0.0008733049],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000804588,"about_ca_system_score_gemma":0.0007284984,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002129182,"about_ca_topic_score_gemma":0.002129819,"domain_scores_codex":[0.999043,0.0003276104,0.00004773188,0.0002063983,0.0002870746,0.00008811025],"domain_scores_gemma":[0.9962978,0.002074517,0.0002271026,0.0006389547,0.0004784318,0.0002832485],"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":[0.001001121,0.0003002927,0.00375371,0.0002267533,0.00009217682,0.0004129916,0.0001747216,0.8588139,0.01121816,0.04801883,0.008416055,0.06757125],"study_design_scores_gemma":[0.00001721353,0.00003674767,0.0001087473,0.000009150447,0.000002710778,0.00002546454,0.00001134288,0.9904185,0.003321912,0.005346082,0.0006945652,0.000007672692],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1542139,0.0002647513,0.8348506,0.0009682633,0.0001942569,0.0001036896,0.0007245493,0.002660033,0.006019993],"genre_scores_gemma":[0.7311571,0.00009029359,0.2620127,0.0004286976,0.00005328607,0.0001577126,0.00159631,0.0006129006,0.003890938],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005790364,"threshold_uncertainty_score":0.01937068,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4200412035","doi":"10.1186/s40537-021-00547-2","title":"Dynamic order Markov model for categorical sequence clustering","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Computer science; Hidden Markov model; Cluster analysis; Pattern recognition (psychology); Markov chain; Markov model; Sequence (biology); Categorical variable; Suffix tree; Data mining; Artificial intelligence; Algorithm; Data structure; Machine learning","authors":[{"name":"Rongbo Chen","is_ca":true},{"name":"Haojun Sun","is_ca":false},{"name":"Lifei Chen","is_ca":false},{"name":"Jianfei Zhang","is_ca":true},{"name":"Shengrui Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1343751196579093,"gpt":0.3358048966184071,"spread":0.2014297769604978,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001415644,0.0007647495,0.001465483,0.001699751,0.0009260456,0.001266636,0.003340684,0.001735629,0.004077865],"category_scores_gemma":[0.004908535,0.0005168018,0.001503816,0.002558939,0.001105,0.0023328,0.00127704,0.002098307,0.001400483],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002209621,"about_ca_system_score_gemma":0.002009471,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01621412,"about_ca_topic_score_gemma":0.01636552,"domain_scores_codex":[0.9985197,0.0003150997,0.00008644565,0.0005578768,0.0003549667,0.0001659719],"domain_scores_gemma":[0.9977367,0.001265651,0.0002950247,0.000250079,0.0003510301,0.000101554],"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":[0.0001252697,0.00007909784,0.003942625,0.0001725525,0.00008611458,0.0002608136,0.0003930384,0.7449486,0.002738634,0.1809092,0.003733318,0.06261088],"study_design_scores_gemma":[0.00000511762,0.00001137073,0.0002146083,0.000005856948,0.000008022006,0.00004007436,0.00001300845,0.9643211,0.0001757384,0.03403002,0.001163073,0.00001200741],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01255827,0.0004349147,0.9834715,0.0003634076,0.00005659505,0.00008828669,0.0006598192,0.000537543,0.001829646],"genre_scores_gemma":[0.6459919,0.001484591,0.3329063,0.0004724401,0.0002221645,0.0008841886,0.004150235,0.0002399829,0.01364826],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01621412,"threshold_uncertainty_score":0.0322395,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4311824268","doi":"10.1186/s40537-022-00667-3","title":"Chromatin state distribution of residue-specific histone acetylation in early myoblast differentiation","year":2022,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Genomics and Chromatin Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada; Southeast University","keywords":"Epigenetics; Chromatin; Histone; Histone acetyltransferase; Acetylation; Computational biology; Biology; Regulation of gene expression; Histone Acetyltransferases; Histone methyltransferase; Cell biology; Genetics; Gene","authors":[{"name":"Yuan Li","is_ca":true},{"name":"Saadia Khilji","is_ca":true},{"name":"Yan Z. Mach","is_ca":true},{"name":"Jihong Chen","is_ca":true},{"name":"Qiao Li","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02572657874015361,"gpt":0.2479546704509838,"spread":0.2222280917108302,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003465576,0.00007779725,0.0001926697,0.000456893,0.000147987,0.0003755533,0.0001256261,0.0001337734,0.0009389437],"category_scores_gemma":[0.0005566693,0.0001251968,0.0001230064,0.0005809706,0.0001803674,0.0001768402,0.0002026236,0.0001993448,0.0002316199],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002222158,"about_ca_system_score_gemma":0.0001348762,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001095352,"about_ca_topic_score_gemma":0.003596508,"domain_scores_codex":[0.9998548,0.00001876673,0.00001036176,0.0000639058,0.00003416699,0.00001792637],"domain_scores_gemma":[0.9997032,0.0001071548,0.00006620394,0.00004244756,0.00005472714,0.00002621526],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0003708183,0.00002156592,0.07783725,0.0001254452,0.00005897963,0.00004977394,0.0002607391,0.001251803,0.9001971,0.0006180815,0.0002113306,0.01899698],"study_design_scores_gemma":[0.000007047756,0.00006855115,0.8535659,0.00001686178,0.00004563831,0.0001774612,0.0001439597,0.006384821,0.1369033,0.000722857,0.001939819,0.00002367716],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9846656,0.0008979877,0.01021357,0.00003676837,0.00000846414,0.00001311487,0.002962265,0.0001207318,0.001081578],"genre_scores_gemma":[0.9930732,0.0002483702,0.003255453,0.00003340365,0.000004901076,0.00001448833,0.002864418,0.00002786331,0.0004780553],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001095352,"threshold_uncertainty_score":0.003141105,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3162192441","doi":"10.1186/s40537-021-00539-2","title":"Integration of image segmentation and fuzzy theory to improve the accuracy of damage detection areas in traffic accidents","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Industrial Vision Systems and Defect Detection","field":"Engineering","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"","keywords":"Machine vision; Computer science; Process (computing); Field (mathematics); Artificial intelligence; Fuzzy logic; Segmentation; Image processing; Computer vision; Machine learning; Image (mathematics)","authors":[{"name":"Majid Amirfakhrian","is_ca":true},{"name":"Mahboub Parhizkar","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04525610434908552,"gpt":0.2922994329218837,"spread":0.2470433285727981,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008057868,0.0004347252,0.0006597306,0.002570213,0.0003534524,0.0006157485,0.0005735357,0.0007598378,0.0008713246],"category_scores_gemma":[0.001747629,0.0003137325,0.0006040354,0.0008649267,0.0003340451,0.0007449996,0.0003182868,0.0004168845,0.0002066141],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005802371,"about_ca_system_score_gemma":0.0006725807,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00731994,"about_ca_topic_score_gemma":0.007129501,"domain_scores_codex":[0.9995669,0.00006039071,0.00003372581,0.00009907329,0.0001829838,0.00005698412],"domain_scores_gemma":[0.9990471,0.0003095713,0.00007593393,0.00005938064,0.0004660175,0.00004196094],"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":[0.0008438475,0.0004556618,0.0136905,0.0002675852,0.0001862017,0.0002532066,0.0002309732,0.1553172,0.09444103,0.002110398,0.002123767,0.7300797],"study_design_scores_gemma":[0.00001315507,0.0001115689,0.005401545,0.0000110757,0.00005416174,0.00008217268,0.00004247604,0.9761013,0.01711741,0.0005431433,0.0005054183,0.00001647848],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3251299,0.001047209,0.6692139,0.0002386498,0.0001655177,0.00008513638,0.0000898316,0.001489455,0.002540483],"genre_scores_gemma":[0.8577056,0.0002499818,0.1409099,0.00005809611,0.00005169166,0.00002544723,0.00008999998,0.00003977221,0.0008695456],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00731994,"threshold_uncertainty_score":0.01455462,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4416285692","doi":"10.1186/s40537-025-01307-2","title":"UniqueNOSD: a novel framework for NoSQL over SQL databases","year":2025,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada; University of Windsor","keywords":"NoSQL; SQL; Scalability; Relational database; Relational database management system; Redundancy (engineering); Consistency (knowledge bases); View; Query by Example","authors":[{"name":"Abdulrauf Gidado","is_ca":true},{"name":"C. I. Ezeife","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1547059754666002,"gpt":0.3786632370180599,"spread":0.2239572615514597,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005514353,0.0009923521,0.001315965,0.001832496,0.001083177,0.004855921,0.005019292,0.00107609,0.01084107],"category_scores_gemma":[0.009844989,0.001022452,0.001479154,0.002495919,0.001403272,0.005651874,0.00554741,0.002990285,0.003225179],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001440024,"about_ca_system_score_gemma":0.004106447,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00369068,"about_ca_topic_score_gemma":0.004967109,"domain_scores_codex":[0.9960912,0.000620369,0.0006467776,0.0005323749,0.001814343,0.0002948414],"domain_scores_gemma":[0.9970136,0.0007903971,0.000228991,0.001025939,0.0007229253,0.0002181015],"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":[0.001127423,0.0005251393,0.005806745,0.00245493,0.0003949628,0.001047348,0.0006259618,0.02511432,0.01682076,0.2888419,0.2013106,0.45593],"study_design_scores_gemma":[0.0004480642,0.0002972017,0.0008950193,0.0003899153,0.0001769046,0.000942726,0.0002133612,0.3206325,0.0241908,0.1512534,0.5002908,0.0002693597],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003554718,0.001092505,0.9321506,0.0007908146,0.0006082603,0.0006480769,0.003277467,0.05198342,0.005894044],"genre_scores_gemma":[0.1033123,0.001717123,0.8634985,0.001437692,0.0005670653,0.001483666,0.01333393,0.005351747,0.009297982],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01084107,"threshold_uncertainty_score":0.03626698,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3169194364","doi":"10.1186/s40537-021-00504-z","title":"The LRA Workbench: an IDE for efficient REST API composition through linked metadata","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Service-Oriented Architecture and Web Services","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Athabasca University; University of Alberta","funders":"","keywords":"Computer science; Workbench; Algorithm; Metadata; Database; Data mining; World Wide Web","authors":[{"name":"Diego Serrano","is_ca":true},{"name":"Eleni Stroulia","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1187062142502125,"gpt":0.3237130792915089,"spread":0.2050068650412964,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009286385,0.002920707,0.001410812,0.002898619,0.001169914,0.005014786,0.00485488,0.001803606,0.02360101],"category_scores_gemma":[0.0109423,0.002181463,0.003272511,0.0008420055,0.001581621,0.006086885,0.006168857,0.004817613,0.01846498],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001149798,"about_ca_system_score_gemma":0.001968123,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002278964,"about_ca_topic_score_gemma":0.002360255,"domain_scores_codex":[0.9958832,0.0012275,0.0005609479,0.0005931742,0.001397078,0.0003381289],"domain_scores_gemma":[0.9927647,0.003553611,0.0005296483,0.001822079,0.0007961298,0.0005339154],"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":[0.00242702,0.002096901,0.006160742,0.001780339,0.0007079341,0.003901012,0.002449355,0.02516067,0.1006652,0.1017068,0.1819907,0.5709534],"study_design_scores_gemma":[0.001174224,0.0005023938,0.002009872,0.0007706443,0.0002173823,0.001191968,0.0004211419,0.4029512,0.1387886,0.04171003,0.4097248,0.0005378505],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002527451,0.00008909134,0.8287463,0.0001364802,0.0001067481,0.0004643264,0.0008687371,0.1628036,0.004257365],"genre_scores_gemma":[0.04525626,0.000352917,0.8604525,0.0006079203,0.0001402338,0.002002684,0.00693554,0.06852409,0.01572781],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02360101,"threshold_uncertainty_score":0.07895327,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4410234140","doi":"10.1186/s40537-025-01141-6","title":"FunDa: scalable serverless data analytics and in situ query processing","year":2025,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Cloud Computing and Resource Management","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada; Mitacs; New Brunswick Innovation Foundation","keywords":"Computer science; Scalability; Analytics; Database; SQL; Data science","authors":[{"name":"Elyes Lounissi","is_ca":false},{"name":"Suvam Kumar Das","is_ca":true},{"name":"Ronnit Peter","is_ca":true},{"name":"Xiaozheng Zhang","is_ca":true},{"name":"Suprio Ray","is_ca":true},{"name":"Lianyin Jia","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1064535844711244,"gpt":0.3074975679074904,"spread":0.201043983436366,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002607703,0.001800982,0.001442319,0.002146227,0.001125815,0.003701626,0.005446384,0.001235367,0.004602379],"category_scores_gemma":[0.007601105,0.0009810929,0.001022244,0.002079678,0.001314087,0.006649278,0.006474651,0.00250075,0.003129408],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001223008,"about_ca_system_score_gemma":0.00268141,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007356219,"about_ca_topic_score_gemma":0.004619215,"domain_scores_codex":[0.9963858,0.0004109697,0.0002396187,0.0007817883,0.001703745,0.0004781465],"domain_scores_gemma":[0.9946401,0.0008405583,0.0003923884,0.002463839,0.000910667,0.0007526478],"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":[0.004148815,0.001984215,0.02025909,0.0011805,0.0008927575,0.001261584,0.001340826,0.03203916,0.08568019,0.03195839,0.272728,0.5465266],"study_design_scores_gemma":[0.0006539915,0.0006110699,0.007178829,0.0001205648,0.0001707045,0.001072526,0.0003990411,0.8201731,0.04508741,0.02677606,0.09742218,0.000334586],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07071552,0.001920205,0.6408348,0.001673863,0.0006128639,0.001193692,0.0042386,0.2652336,0.0135768],"genre_scores_gemma":[0.6097093,0.0008814684,0.3601191,0.001488177,0.0003990352,0.0006596484,0.01200152,0.004827154,0.009914629],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007356219,"threshold_uncertainty_score":0.01539648,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3124386847","doi":"10.1186/s40537-021-00418-w","title":"CaReAl: capturing read alignments in a BAM file rapidly and conveniently","year":2021,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Princess Margaret Cancer Centre; University Health Network","funders":"National Institute of Food and Drug Safety Evaluation; Ministry of Food and Drug Safety","keywords":"Computer science; Snapshot (computer storage); Visualization; Multithreading; Data mining; Database; Operating system; Thread (computing)","authors":[{"name":"Yoomi Park","is_ca":false},{"name":"Heewon Seo","is_ca":true},{"name":"Kyung Hun Yoo","is_ca":false},{"name":"Ju Han Kim","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05366287355894767,"gpt":0.2675212990260544,"spread":0.2138584254671067,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002654247,0.00242273,0.0008762137,0.003516215,0.0008495378,0.002258352,0.00240295,0.0009601822,0.01887442],"category_scores_gemma":[0.008849668,0.000929145,0.0008615281,0.002289946,0.0004476477,0.002127512,0.001812407,0.001974377,0.008550096],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005908191,"about_ca_system_score_gemma":0.0009052908,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001883042,"about_ca_topic_score_gemma":0.002902404,"domain_scores_codex":[0.9983347,0.0003060061,0.0001744633,0.0004960835,0.0005269482,0.0001617972],"domain_scores_gemma":[0.9952251,0.002094631,0.0005169916,0.0009290535,0.0009204347,0.0003138262],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003211828,0.0002549949,0.007137129,0.002530502,0.0003098794,0.001100464,0.00120317,0.002307074,0.2816453,0.005059607,0.2215557,0.4736844],"study_design_scores_gemma":[0.0003907975,0.0004206647,0.0210196,0.0005398447,0.0001750817,0.001512982,0.0005023857,0.06023245,0.6427234,0.007123966,0.2647611,0.000597866],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.02545479,0.0006899859,0.5781906,0.0004439808,0.0004453318,0.0007276942,0.03199773,0.3568631,0.005186873],"genre_scores_gemma":[0.0891458,0.0005042485,0.8436969,0.0006315493,0.0001461562,0.001345505,0.03181997,0.02704431,0.005665501],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.01887442,"threshold_uncertainty_score":0.06314117,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4409069510","doi":"10.1186/s40537-025-01122-9","title":"SPINEX-anomaly: similarity-based predictions with explainable neighbors exploration for anomaly and outlier detection","year":2025,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba","funders":"Clemson University","keywords":"Anomaly detection; Anomaly (physics); Outlier; Computer science; Similarity (geometry); Data mining; Computational Science and Engineering; Artificial intelligence; Pattern recognition (psychology); Machine learning; Image (mathematics)","authors":[{"name":"M.Z. Naser","is_ca":false},{"name":"Ahmad Z. Naser","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06406157203631375,"gpt":0.2900970408861074,"spread":0.2260354688497937,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001965711,0.001420149,0.001631431,0.002095006,0.0007181074,0.001373551,0.002668122,0.001172824,0.001568],"category_scores_gemma":[0.009443905,0.0004068364,0.001096065,0.001717665,0.0009173091,0.002439733,0.002775323,0.002185114,0.0006450041],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006884231,"about_ca_system_score_gemma":0.001692352,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004102705,"about_ca_topic_score_gemma":0.004192894,"domain_scores_codex":[0.9984182,0.000372341,0.00009774135,0.0003913969,0.0005928412,0.000127461],"domain_scores_gemma":[0.9962633,0.001913113,0.0004424176,0.0005493287,0.0006588022,0.0001729982],"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":[0.0006914415,0.0004161982,0.02720182,0.0002783998,0.0003501512,0.0004400358,0.0005039947,0.4250292,0.007817154,0.01425831,0.01253133,0.510482],"study_design_scores_gemma":[0.0000189383,0.00007731171,0.000569171,0.000008653948,0.00001222725,0.00007599459,0.00004247712,0.9906471,0.001436032,0.00624574,0.0008551528,0.0000110238],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0591675,0.0006511876,0.9318586,0.0004662693,0.0001075089,0.0001709948,0.0005094618,0.006047404,0.001020992],"genre_scores_gemma":[0.6202494,0.0002618174,0.3752018,0.0002589145,0.0001319623,0.0002227988,0.00180793,0.0002901884,0.001575151],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004102705,"threshold_uncertainty_score":0.01039582,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2806417635","doi":"10.1186/s40537-018-0128-5","title":"A non-parametric maximum for number of selected features: objective optima for FDR and significance threshold with application to ordinal survey analysis","year":2018,"lang":"en","type":"article","venue":"Journal Of Big Data","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Capilano University; Simon Fraser University","funders":"","keywords":"False discovery rate; Statistical hypothesis testing; Mathematics; False positives and false negatives; Statistics; Multiple comparisons problem; Parametric statistics; Ordinal data; False positive paradox; Computer science; Nonparametric statistics; Set (abstract data type); Data mining; Artificial intelligence; Pattern recognition (psychology)","authors":[{"name":"Amir Hassan Ghaseminejad Tafreshi","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1602111434238936,"gpt":0.446176169199242,"spread":0.2859650257753483,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.07064896,0.001492707,0.002924729,0.003497943,0.001634517,0.00282081,0.003971596,0.003318716,0.001880001],"category_scores_gemma":[0.2232967,0.0008263541,0.002251651,0.003162873,0.00329288,0.003567619,0.004299972,0.004376947,0.000518446],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001856504,"about_ca_system_score_gemma":0.003139942,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009663551,"about_ca_topic_score_gemma":0.0008415041,"domain_scores_codex":[0.9486678,0.03846556,0.002071861,0.003638163,0.006159208,0.0009975386],"domain_scores_gemma":[0.7826437,0.1890058,0.006815595,0.00933656,0.01109314,0.001105149],"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":[0.002443468,0.0004792387,0.02249555,0.002703512,0.0009524747,0.001201168,0.001804106,0.1883449,0.01539701,0.2009481,0.006706838,0.5565237],"study_design_scores_gemma":[0.0002557526,0.0009772968,0.006753265,0.000536715,0.0003156867,0.001297343,0.000512055,0.7537277,0.01291115,0.2133578,0.009149825,0.0002053523],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004862602,0.0003908388,0.9934759,0.0003449608,0.00003958646,0.0001282486,0.00006349966,0.0002153684,0.0004791584],"genre_scores_gemma":[0.1099671,0.000227293,0.8878526,0.0001793983,0.00008462483,0.0008357398,0.0001187981,0.0001710963,0.0005634704],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.07064896,"threshold_uncertainty_score":0.3736319,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}