{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":10,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":10,"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":"09ce87fd00b5","filters":{"venue":"Big Data Research"}},"results":[{"id":"W2922919311","doi":"10.1016/j.bdr.2019.03.001","title":"Systematic Review of the Literature on Big Data in the Transportation Domain: Concepts and Applications","year":2019,"lang":"en","type":"article","venue":"Big Data Research","topic":"Traffic Prediction and Management Techniques","field":"Engineering","cited_by":124,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"MacEwan University","funders":"MacEwan University; University of Otago; U.S. Department of Transportation","keywords":"Big data; Data science; Computer science; Crowdsourcing; Domain (mathematical analysis); Analytics; SAFER; Intelligent transportation system; Data analysis; Software analytics; Software; World Wide Web; Computer security; Software development; Data mining; Engineering; Transport engineering; Component-based software engineering","authors":[{"name":"Alex Neilson","is_ca":true},{"name":"Indratmo","is_ca":true},{"name":"Ben Kei Daniel","is_ca":false},{"name":"Stevanus A. Tjandra","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1301934818627772,"gpt":0.365128430725414,"spread":0.2349349488626367,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01879782,0.00145846,0.005141613,0.02167013,0.0009981104,0.003725937,0.001685278,0.00183264,0.003249742],"category_scores_gemma":[0.09687281,0.0009955965,0.005649026,0.01949172,0.001613281,0.003327907,0.002444016,0.001845373,0.0003439219],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003650627,"about_ca_system_score_gemma":0.03168301,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01087729,"about_ca_topic_score_gemma":0.0385879,"domain_scores_codex":[0.9846361,0.006219184,0.004879897,0.001122237,0.002846264,0.0002963347],"domain_scores_gemma":[0.8500845,0.1215752,0.01362332,0.002879743,0.01067779,0.001159491],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.0002130149,0.00006515041,0.003205079,0.8436979,0.007465216,0.0001681443,0.0006375897,0.0005096336,0.0003061588,0.002398299,0.00841056,0.1329233],"study_design_scores_gemma":[0.0001527294,0.0002413067,0.008274613,0.8975039,0.02461647,0.0003581403,0.001059783,0.0003165832,0.0003370928,0.003858237,0.06319834,0.00008277817],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00192295,0.9919859,0.001563095,0.001433258,0.000357405,0.0004254526,0.00175713,0.00001646473,0.0005382789],"genre_scores_gemma":[0.02179893,0.9695249,0.004592767,0.001586857,0.0002952765,0.001002189,0.001035021,0.00001214602,0.000151889],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.02167013,"threshold_uncertainty_score":0.09941363,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3113091126","doi":"10.1016/j.bdr.2020.100173","title":"Agreeing to Disagree: Choosing Among Eight Topic-Modeling Methods","year":2020,"lang":"en","type":"article","venue":"Big Data Research","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":31,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University; University of British Columbia","funders":"Hong Kong Polytechnic University","keywords":"Latent Dirichlet allocation; Topic model; Computer science; Data science; Artificial intelligence; Coherence (philosophical gambling strategy); Field (mathematics); Context (archaeology); Machine learning; Management science; Natural language processing","authors":[{"name":"Qiang Fu","is_ca":true},{"name":"Yufan Zhuang","is_ca":false},{"name":"Jiaxin Gu","is_ca":true},{"name":"Yushu Zhu","is_ca":true},{"name":"Xin Guo","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.8645336607084751,"gpt":0.6295130273678069,"spread":0.2350206333406683,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.1976716,0.002535491,0.003023694,0.006498765,0.004347725,0.007980894,0.005831443,0.005154657,0.006813993],"category_scores_gemma":[0.3329974,0.001635352,0.004478248,0.004767443,0.003544451,0.01111379,0.007425272,0.007263245,0.002170931],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001750238,"about_ca_system_score_gemma":0.003889166,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00136597,"about_ca_topic_score_gemma":0.004318592,"domain_scores_codex":[0.8347772,0.1381306,0.007954801,0.00594509,0.01091581,0.002276591],"domain_scores_gemma":[0.3774642,0.5812154,0.008933019,0.01187027,0.01562411,0.004893096],"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":[0.01206268,0.002374437,0.08158297,0.004764759,0.004213657,0.0002374645,0.02465527,0.008747776,0.00364413,0.04218261,0.02420812,0.7913261],"study_design_scores_gemma":[0.006882404,0.002645764,0.04922847,0.004978783,0.006237485,0.0008355011,0.0321158,0.4033126,0.01330726,0.4376624,0.04152385,0.00126977],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2081795,0.004004317,0.7457227,0.01238216,0.001131329,0.006634318,0.00131667,0.001910983,0.01871799],"genre_scores_gemma":[0.440204,0.00103177,0.5450583,0.001797486,0.0004251892,0.007083504,0.001521211,0.0005649134,0.002313689],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1976716,"threshold_uncertainty_score":0.9894137,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2901153708","doi":"10.1016/j.bdr.2018.11.001","title":"Label-Aware Distributed Ensemble Learning: A Simplified Distributed Classifier Training Model for Big Data","year":2018,"lang":"en","type":"article","venue":"Big Data Research","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"IBM (Canada); Queen's University","funders":"","keywords":"Computer science; Classifier (UML); Big data; Data mining; Machine learning; SPARK (programming language); Artificial intelligence","authors":[{"name":"Shadi Khalifa","is_ca":true},{"name":"Patrick Martin","is_ca":true},{"name":"Rebecca C. Young","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.72474036327115,"gpt":0.4627651775700344,"spread":0.2619751857011156,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00200847,0.0008986053,0.001860843,0.0005914084,0.00105826,0.001520215,0.003904176,0.001561925,0.003182174],"category_scores_gemma":[0.005247181,0.000675851,0.000990986,0.001097981,0.0005444711,0.003694092,0.002691079,0.002705659,0.001638658],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007606695,"about_ca_system_score_gemma":0.001878547,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005877874,"about_ca_topic_score_gemma":0.0122133,"domain_scores_codex":[0.9988305,0.0002542454,0.00005976758,0.0003476012,0.0003697744,0.0001380751],"domain_scores_gemma":[0.9974636,0.000678102,0.0001021047,0.0009112055,0.000697122,0.0001477547],"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":[0.0005932327,0.0005234783,0.002729523,0.0000774855,0.0001730104,0.0001373586,0.0001586918,0.5143271,0.007383924,0.009248092,0.008701949,0.4559462],"study_design_scores_gemma":[0.000008284867,0.0000203189,0.00007985581,0.00000181692,0.00001156286,0.00001374112,0.000007583071,0.9949493,0.0006333975,0.003891747,0.0003780884,0.000004185889],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009580781,0.0001156764,0.9878194,0.000143673,0.0000706374,0.00005822728,0.00008269201,0.001491165,0.000637796],"genre_scores_gemma":[0.448376,0.0002421932,0.542465,0.0003143918,0.0003056724,0.000380343,0.0007679663,0.0005217772,0.00662671],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005877874,"threshold_uncertainty_score":0.01168734,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4295709232","doi":"10.1016/j.bdr.2022.100348","title":"Properties and Performance of the ABCDe Random Graph Model with Community Structure","year":2022,"lang":"en","type":"article","venue":"Big Data Research","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":19,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Computer science; Benchmark (surveying); Graph; Random graph; Generator (circuit theory); Community structure; Theoretical computer science; Algorithm; Power (physics); Mathematics","authors":[{"name":"Bogumił Kamiński","is_ca":false},{"name":"Tomasz Olczak","is_ca":false},{"name":"Bartosz Pankratz","is_ca":false},{"name":"Paweł Prałat","is_ca":true},{"name":"François Théberge","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2926327152332855,"gpt":0.3403536589814936,"spread":0.04772094374820812,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006241729,0.0007606768,0.001002968,0.001544934,0.0009170299,0.001620361,0.002126688,0.001544134,0.001630207],"category_scores_gemma":[0.03727945,0.0005292277,0.0006605778,0.001180396,0.001304586,0.00315817,0.001261581,0.001517607,0.0006041048],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001819028,"about_ca_system_score_gemma":0.00149735,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008540486,"about_ca_topic_score_gemma":0.006863521,"domain_scores_codex":[0.997387,0.001054352,0.0001127416,0.0006078207,0.0006402305,0.000197817],"domain_scores_gemma":[0.9777294,0.01395741,0.001201628,0.003943409,0.002533048,0.0006351136],"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":[0.0006084232,0.0003075543,0.006768093,0.0002081507,0.0001314176,0.0001122484,0.0001238033,0.9331537,0.004031784,0.01720636,0.006856188,0.03049229],"study_design_scores_gemma":[0.00003665467,0.00005061119,0.000272131,0.000005364736,0.000006638408,0.00003607494,0.00001288849,0.9926122,0.001041559,0.005539734,0.0003793446,0.000006653322],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6701952,0.001178092,0.3084263,0.002024008,0.000212764,0.0003315272,0.001896902,0.006195643,0.009539519],"genre_scores_gemma":[0.8633912,0.0002638275,0.1302892,0.00029935,0.00005203333,0.0002898938,0.003148409,0.0005034576,0.00176248],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.008540486,"threshold_uncertainty_score":0.03300977,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3208577093","doi":"10.1016/j.bdr.2021.100275","title":"Epidemic Spreading in Trajectory Networks","year":2021,"lang":"en","type":"article","venue":"Big Data Research","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":17,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Outbreak; Psychological intervention; Intervention (counseling); Contact tracing; Coronavirus disease 2019 (COVID-19); Disease; Epidemic model; Infectious disease (medical specialty); Big data; Data science; Risk analysis (engineering); Computer security; Medicine; Environmental health; Virology; Data mining; Population","authors":[{"name":"Tilemachos Pechlivanoglou","is_ca":true},{"name":"Jing Li","is_ca":true},{"name":"J Sun","is_ca":true},{"name":"Farzaneh Heidari","is_ca":true},{"name":"Manos Papagelis","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.9041873969734017,"gpt":0.5915173642557057,"spread":0.312670032717696,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003206594,0.0006868368,0.001164794,0.00390471,0.001790336,0.004005205,0.001644405,0.002131603,0.005579627],"category_scores_gemma":[0.0414417,0.0009005244,0.001158188,0.003253558,0.002702962,0.007529044,0.002998753,0.002339317,0.0007191086],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002022617,"about_ca_system_score_gemma":0.0009783566,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008227238,"about_ca_topic_score_gemma":0.00400298,"domain_scores_codex":[0.9981687,0.0008956382,0.00008751082,0.00039404,0.0002829144,0.0001711507],"domain_scores_gemma":[0.966249,0.0253283,0.003907161,0.001731667,0.001747314,0.001036485],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001130074,0.00006080835,0.008531292,0.0001691187,0.00009829074,0.0002813821,0.001250532,0.1593873,0.0007445651,0.8091912,0.004239142,0.01593337],"study_design_scores_gemma":[0.00002789946,0.00002415184,0.00130266,0.00004928957,0.00003422407,0.0001782278,0.000343807,0.5519982,0.0002507448,0.4428458,0.002923741,0.00002132055],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3595092,0.002120124,0.6133965,0.006200924,0.0001496449,0.0001897143,0.001228673,0.0004261077,0.01677905],"genre_scores_gemma":[0.9568819,0.002349357,0.02696214,0.000282767,0.0002109671,0.0002098582,0.0009648397,0.00009654855,0.01204158],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008227238,"threshold_uncertainty_score":0.01866567,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3145041349","doi":"10.1016/j.bdr.2021.100254","title":"Scaling the Growing Neural Gas for Visual Cluster Analysis","year":2021,"lang":"en","type":"article","venue":"Big Data Research","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Clutter; Visualization; Neural gas; Big data; Cluster (spacecraft); Artificial intelligence; Unsupervised learning; Graph; Scalability; Artificial neural network; Topology (electrical circuits); Machine learning; Pattern recognition (psychology); Theoretical computer science; Data mining; Recurrent neural network; Database","authors":[{"name":"Elio Ventocilla","is_ca":false},{"name":"Rafael Martins","is_ca":false},{"name":"Fernando V. Paulovich","is_ca":true},{"name":"Maria Riveiro","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4006687696177129,"gpt":0.4578600640552944,"spread":0.05719129443758159,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001062622,0.00117488,0.0007138008,0.001523257,0.0006399537,0.001578096,0.001492533,0.001103227,0.003432007],"category_scores_gemma":[0.007950963,0.00046968,0.0009206284,0.001321822,0.0008497018,0.002636793,0.001563618,0.001331163,0.0009216772],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001356291,"about_ca_system_score_gemma":0.0008654885,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01007719,"about_ca_topic_score_gemma":0.008350087,"domain_scores_codex":[0.9994228,0.0001774257,0.00002355467,0.0001485133,0.0001781026,0.00004959244],"domain_scores_gemma":[0.9982367,0.0008786274,0.0001234293,0.0002779302,0.0004098831,0.00007358667],"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":[0.0001683516,0.00005544276,0.001733132,0.0001549897,0.00005791349,0.0001438063,0.0002873957,0.8251445,0.006881826,0.01900425,0.006239167,0.1401293],"study_design_scores_gemma":[0.000004907086,0.000008829605,0.0001235947,0.000007505154,0.000003946031,0.00002048677,0.00002716037,0.989614,0.001132252,0.007952059,0.001098108,0.000007278491],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04925209,0.000722344,0.9381126,0.0005897175,0.000159042,0.0001113458,0.0003605224,0.004463575,0.006228889],"genre_scores_gemma":[0.5120638,0.0005591228,0.4816382,0.0003500659,0.00007346602,0.0002278736,0.0008245702,0.001179941,0.003082958],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01007719,"threshold_uncertainty_score":0.02003706,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3170000029","doi":"10.1016/j.bdr.2021.100239","title":"ExplorerTree: A Focus+Context Exploration Approach for 2D Embeddings","year":2021,"lang":"en","type":"article","venue":"Big Data Research","topic":"Data Visualization and Analytics","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Computer science; Focus (optics); Artificial intelligence; Context (archaeology); Clutter; Convolutional neural network; Dimensionality reduction; Visualization; Machine learning; Pattern recognition (psychology); Data mining; Radar","authors":[{"name":"Wilson E. Marcílio-Jr","is_ca":false},{"name":"Danilo Medeiros Eler","is_ca":false},{"name":"Fernando V. Paulovich","is_ca":true},{"name":"José F. Rodrigues","is_ca":false},{"name":"Almir Olivette Artero","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6266507062173049,"gpt":0.4639213162888603,"spread":0.1627293899284446,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009787785,0.002712275,0.001738524,0.003216688,0.0009824823,0.003619527,0.002749282,0.00165537,0.01598017],"category_scores_gemma":[0.005527802,0.001219595,0.002568166,0.002726639,0.0007655082,0.004205901,0.008339535,0.002143735,0.003596112],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000486586,"about_ca_system_score_gemma":0.0009897354,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00359972,"about_ca_topic_score_gemma":0.01229847,"domain_scores_codex":[0.9991716,0.0001929339,0.00004636513,0.0001767602,0.0003292136,0.00008312021],"domain_scores_gemma":[0.9984365,0.0008341026,0.00007351385,0.0003371729,0.0002135884,0.0001049508],"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":[0.0008793224,0.000367788,0.004074796,0.002710268,0.000477304,0.000781844,0.004699807,0.04536375,0.04039535,0.05763375,0.05950159,0.7831145],"study_design_scores_gemma":[0.0001850189,0.0002918825,0.001204846,0.0003517915,0.0001897784,0.0009878624,0.001801825,0.7472759,0.03003297,0.1075935,0.1098773,0.0002072382],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005986065,0.0004786288,0.9697562,0.000199841,0.00008409636,0.0001675563,0.002129951,0.0192463,0.001951344],"genre_scores_gemma":[0.07052478,0.0005248902,0.9181077,0.0001912028,0.00003901693,0.0003983455,0.003588459,0.00396565,0.002659967],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01598017,"threshold_uncertainty_score":0.05345905,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413323615","doi":"10.1016/j.bdr.2025.100555","title":"Explainable malware detection through integrated graph reduction and learning techniques","year":2025,"lang":"en","type":"article","venue":"Big Data Research","topic":"Advanced Malware Detection Techniques","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of New Brunswick","funders":"","keywords":"Computer science; Malware; Reduction (mathematics); Big data; Graph; Artificial intelligence; Data science; Machine learning; Deep learning; Computer security; Theoretical computer science; Data mining; Mathematics","authors":[{"name":"Hesamodin Mohammadian","is_ca":true},{"name":"Griffin Higgins","is_ca":true},{"name":"Samuel Ansong","is_ca":true},{"name":"Roozbeh Razavi‐Far","is_ca":true},{"name":"Ali A. Ghorbani","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1808340169138115,"gpt":0.4031757345333891,"spread":0.2223417176195776,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004633247,0.001023012,0.000511776,0.002356116,0.0004642303,0.0006415534,0.0008846417,0.0007240467,0.001322038],"category_scores_gemma":[0.003259009,0.0003758807,0.00127618,0.0008375891,0.0006980544,0.001372894,0.0009646087,0.001248008,0.0002998075],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008082842,"about_ca_system_score_gemma":0.00085522,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005317615,"about_ca_topic_score_gemma":0.009556659,"domain_scores_codex":[0.9995511,0.0001232467,0.00001541866,0.0001319505,0.0001358303,0.00004244515],"domain_scores_gemma":[0.9985252,0.0008671543,0.0001777256,0.00022665,0.0001696091,0.0000336994],"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":[0.0001174021,0.0002274563,0.007354483,0.0001871537,0.0001488459,0.0004009767,0.0003054022,0.5718219,0.01427594,0.02755925,0.004127813,0.3734734],"study_design_scores_gemma":[0.00000470113,0.00001774323,0.0005745122,0.000007657289,0.00001989621,0.00004812337,0.00002614689,0.9753627,0.002077518,0.02084817,0.001006966,0.000005886414],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04961557,0.0002230647,0.9463026,0.0004499898,0.00002034182,0.00007497349,0.0002176115,0.002113438,0.000982393],"genre_scores_gemma":[0.5392759,0.0004229395,0.4553617,0.0002144121,0.00005661245,0.0001602085,0.001541393,0.000344258,0.002622621],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005317615,"threshold_uncertainty_score":0.01057333,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4405429927","doi":"10.1016/j.bdr.2024.100504","title":"Principal component analysis of multivariate spatial functional data","year":2024,"lang":"en","type":"article","venue":"Big Data Research","topic":"Soil Geostatistics and Mapping","field":"Environmental Science","cited_by":6,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Principal component analysis; Multivariate statistics; Computer science; Big data; Multivariate analysis; Component (thermodynamics); Data mining; Artificial intelligence; Pattern recognition (psychology); Machine learning","authors":[{"name":"Si-Ahmed Idris","is_ca":false},{"name":"Leila Hamdad","is_ca":false},{"name":"Christelle Judith Agonkoui","is_ca":false},{"name":"Yoba Kande","is_ca":false},{"name":"Sophie Dabo‐Niang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4518594465811662,"gpt":0.4224598545003331,"spread":0.02939959208083309,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002363251,0.001303502,0.0008747523,0.003302093,0.0005746002,0.001873667,0.0006943041,0.0003789501,0.002131748],"category_scores_gemma":[0.01130898,0.0003921904,0.001610677,0.004391936,0.0007531873,0.001342864,0.001091686,0.001191622,0.0008860311],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005238044,"about_ca_system_score_gemma":0.001702462,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00567335,"about_ca_topic_score_gemma":0.003915562,"domain_scores_codex":[0.9977526,0.001036959,0.0001323967,0.0003317051,0.0005860321,0.0001603049],"domain_scores_gemma":[0.9944165,0.003260096,0.0003433363,0.0009685708,0.0008992756,0.0001123396],"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":[0.000296377,0.0002102155,0.01355157,0.0006139387,0.0008127319,0.0001860583,0.0004656932,0.1391906,0.01128322,0.08306108,0.01604392,0.7342846],"study_design_scores_gemma":[0.00002058718,0.0000527418,0.01437199,0.00005817233,0.0001231057,0.0001353694,0.0001709007,0.8836678,0.002655145,0.08946229,0.009224762,0.00005720195],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01909834,0.0003316572,0.9771892,0.0002292288,0.00008278345,0.00006001593,0.00108266,0.00115208,0.0007741005],"genre_scores_gemma":[0.4493339,0.001331287,0.5368565,0.0001303415,0.0002641492,0.0005478975,0.007079534,0.001102315,0.003354097],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00567335,"threshold_uncertainty_score":0.0124982,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2005161541","doi":"10.1016/j.bdr.2014.08.001","title":"Special Issue on Scalable Computing for Big Data","year":2014,"lang":"en","type":"article","venue":"Big Data Research","topic":"Cloud Computing and Resource Management","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"St. Francis Xavier University","funders":"","keywords":"Big data; Computer science; Scalability; Data science; Distributed computing; Data mining; Database","authors":[{"name":"Laurence T. Yang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3942306666903458,"gpt":0.4020958052569457,"spread":0.007865138566599927,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003650339,0.003004613,0.003572909,0.004104987,0.002483576,0.007145049,0.002872156,0.005650931,0.09281572],"category_scores_gemma":[0.006795568,0.0009460215,0.002100106,0.002624299,0.001403583,0.00518121,0.002530548,0.005556785,0.03007041],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001864749,"about_ca_system_score_gemma":0.002239535,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006357239,"about_ca_topic_score_gemma":0.001712085,"domain_scores_codex":[0.9975287,0.0002517948,0.0002343981,0.000442821,0.001213292,0.0003289806],"domain_scores_gemma":[0.9883478,0.00306444,0.0008409588,0.0009065921,0.004105028,0.00273512],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00008077943,0.00005707306,0.0001657216,0.0004813618,0.00003826701,0.0001381328,0.00001043674,0.0001627379,0.0004490696,0.002079578,0.9746166,0.02172025],"study_design_scores_gemma":[0.00003066276,0.00008908105,0.0008556288,0.0001818174,0.0000367404,0.0002308477,0.00001998734,0.0006800958,0.0003008628,0.003394213,0.9941597,0.00002041005],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.0005310235,0.01451121,0.001513961,0.02490536,0.9231021,0.0001338863,0.0003465223,0.0002911709,0.03466485],"genre_scores_gemma":[0.002287214,0.01237753,0.0005482105,0.005688942,0.9096676,0.00008920771,0.0006277782,0.000354109,0.06835936],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.09281572,"threshold_uncertainty_score":0.3104995,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}