{"meta":{"query_hash":"09ce87fd00b5","filters":{"venue":"Big Data Research"},"cohort_total":10,"direct_labels_cover":0,"predictions_cover":10,"exported":10,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/09ce87fd00b5","api":"https://metacan.xera.ac/api/v1/cohort?venue=Big+Data+Research"},"results":[{"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"St. Francis Xavier University","funders":"","keywords":"Big data; Computer science; Scalability; Data science; Distributed computing; Data mining; Database","score_opus":0.3942306666903458,"score_gpt":0.4020958052569457,"score_spread":0.007865138566599927,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2005161541","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00053102354,0.014511212,0.0015139614,0.024905363,0.9231021,0.00013388632,0.00034652225,0.00029117093,0.034664854],"genre_scores_gemma":[0.0022872144,0.01237753,0.0005482105,0.005688942,0.9096676,0.00008920771,0.00062777824,0.000354109,0.06835936],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99752873,0.00025179476,0.0002343981,0.00044282098,0.0012132919,0.0003289806],"domain_scores_gemma":[0.9883478,0.0030644399,0.0008409588,0.0009065921,0.004105028,0.00273512],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036503393,0.003004613,0.0035729087,0.004104987,0.0024835758,0.0071450486,0.0028721564,0.005650931,0.09281572],"category_scores_gemma":[0.0067955684,0.00094602146,0.0021001056,0.002624299,0.0014035833,0.0051812096,0.002530548,0.0055567846,0.030070413],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000080779435,0.00005707306,0.00016572159,0.0004813618,0.000038267008,0.0001381328,0.000010436745,0.00016273787,0.00044906957,0.0020795781,0.9746166,0.02172025],"study_design_scores_gemma":[0.00003066276,0.00008908105,0.00085562875,0.00018181736,0.000036740395,0.00023084765,0.000019987336,0.0006800958,0.0003008628,0.0033942133,0.9941597,0.000020410047],"about_ca_topic_score_codex":0.00063572393,"about_ca_topic_score_gemma":0.0017120854,"teacher_disagreement_score":0.09281572,"about_ca_system_score_codex":0.0018647485,"about_ca_system_score_gemma":0.0022395349,"threshold_uncertainty_score":0.31049955},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.72474036327115,"score_gpt":0.4627651775700344,"score_spread":0.2619751857011156,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2901153708","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009580781,0.00011567636,0.98781943,0.000143673,0.000070637405,0.00005822728,0.000082692015,0.001491165,0.00063779595],"genre_scores_gemma":[0.44837597,0.00024219316,0.54246503,0.00031439177,0.00030567244,0.00038034297,0.00076796627,0.0005217772,0.0066267103],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988305,0.00025424545,0.000059767575,0.0003476012,0.00036977438,0.00013807512],"domain_scores_gemma":[0.9974636,0.000678102,0.00010210474,0.00091120554,0.000697122,0.00014775473],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0020084698,0.00089860527,0.0018608432,0.0005914084,0.0010582599,0.0015202146,0.0039041762,0.0015619247,0.0031821737],"category_scores_gemma":[0.005247181,0.00067585096,0.000990986,0.0010979814,0.00054447114,0.0036940917,0.0026910794,0.0027056593,0.0016386579],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005932327,0.00052347826,0.0027295235,0.0000774855,0.00017301038,0.00013735858,0.00015869181,0.5143271,0.007383924,0.009248092,0.008701949,0.45594618],"study_design_scores_gemma":[0.000008284867,0.000020318896,0.00007985581,0.0000018169195,0.000011562858,0.000013741123,0.0000075830712,0.99494934,0.00063339755,0.0038917465,0.0003780884,0.000004185889],"about_ca_topic_score_codex":0.0058778743,"about_ca_topic_score_gemma":0.012213304,"teacher_disagreement_score":0.0058778743,"about_ca_system_score_codex":0.0007606695,"about_ca_system_score_gemma":0.001878547,"threshold_uncertainty_score":0.011687338},"labels":[],"label_agreement":null},{"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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","score_opus":0.13019348186277724,"score_gpt":0.36512843072541395,"score_spread":0.23493494886263672,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2922919311","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0019229504,0.9919859,0.0015630949,0.001433258,0.00035740496,0.00042545257,0.0017571297,0.000016464728,0.0005382789],"genre_scores_gemma":[0.021798935,0.9695249,0.0045927665,0.0015868575,0.00029527652,0.0010021895,0.0010350206,0.000012146023,0.00015188896],"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","domain_scores_codex":[0.98463607,0.006219184,0.0048798965,0.0011222371,0.0028462636,0.00029633474],"domain_scores_gemma":[0.8500845,0.12157521,0.0136233205,0.0028797432,0.010677791,0.0011594907],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.018797824,0.0014584597,0.005141613,0.021670135,0.0009981104,0.0037259366,0.001685278,0.0018326403,0.0032497423],"category_scores_gemma":[0.09687281,0.0009955965,0.0056490265,0.019491717,0.0016132809,0.003327907,0.0024440156,0.0018453732,0.00034392189],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00021301494,0.00006515041,0.0032050791,0.8436979,0.0074652163,0.0001681443,0.00063758966,0.0005096336,0.00030615876,0.002398299,0.00841056,0.13292329],"study_design_scores_gemma":[0.00015272936,0.00024130668,0.008274613,0.8975039,0.024616469,0.00035814027,0.0010597834,0.00031658317,0.0003370928,0.0038582368,0.063198335,0.00008277817],"about_ca_topic_score_codex":0.010877294,"about_ca_topic_score_gemma":0.038587898,"teacher_disagreement_score":0.021670135,"about_ca_system_score_codex":0.0036506266,"about_ca_system_score_gemma":0.031683005,"threshold_uncertainty_score":0.09941363},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.8645336607084751,"score_gpt":0.6295130273678069,"score_spread":0.23502063334066825,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3113091126","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.20817952,0.004004317,0.7457227,0.012382163,0.0011313291,0.006634318,0.0013166703,0.0019109832,0.018717991],"genre_scores_gemma":[0.44020402,0.0010317698,0.54505825,0.0017974864,0.0004251892,0.0070835035,0.001521211,0.0005649134,0.0023136889],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.8347772,0.13813059,0.007954801,0.0059450897,0.010915814,0.0022765908],"domain_scores_gemma":[0.37746418,0.5812154,0.008933019,0.01187027,0.015624114,0.0048930957],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.19767164,0.002535491,0.0030236938,0.006498765,0.004347725,0.007980894,0.005831443,0.005154657,0.0068139927],"category_scores_gemma":[0.33299738,0.0016353521,0.004478248,0.0047674426,0.0035444514,0.011113788,0.007425272,0.007263245,0.002170931],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.012062685,0.0023744374,0.08158297,0.0047647594,0.0042136568,0.00023746447,0.024655268,0.008747776,0.0036441304,0.042182613,0.024208115,0.7913261],"study_design_scores_gemma":[0.0068824044,0.0026457645,0.049228467,0.0049787834,0.0062374845,0.00083550107,0.0321158,0.4033126,0.013307264,0.43766236,0.041523848,0.0012697695],"about_ca_topic_score_codex":0.0013659702,"about_ca_topic_score_gemma":0.004318592,"teacher_disagreement_score":0.19767164,"about_ca_system_score_codex":0.0017502378,"about_ca_system_score_gemma":0.0038891658,"threshold_uncertainty_score":0.98941374},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.40066876961771286,"score_gpt":0.45786006405529445,"score_spread":0.057191294437581586,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3145041349","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04925209,0.000722344,0.93811256,0.00058971747,0.00015904197,0.00011134582,0.00036052242,0.004463575,0.0062288893],"genre_scores_gemma":[0.5120638,0.0005591228,0.48163822,0.00035006588,0.00007346602,0.00022787358,0.0008245702,0.001179941,0.0030829576],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994228,0.0001774257,0.000023554667,0.00014851331,0.00017810262,0.000049592443],"domain_scores_gemma":[0.99823666,0.0008786274,0.00012342927,0.00027793017,0.00040988313,0.00007358667],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010626216,0.0011748798,0.00071380084,0.0015232568,0.0006399537,0.001578096,0.0014925331,0.0011032269,0.0034320073],"category_scores_gemma":[0.007950963,0.00046968,0.0009206284,0.0013218222,0.00084970176,0.0026367928,0.0015636181,0.0013311635,0.0009216772],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0001683516,0.000055442757,0.0017331318,0.00015498968,0.00005791349,0.00014380633,0.00028739567,0.8251445,0.006881826,0.019004252,0.0062391674,0.1401293],"study_design_scores_gemma":[0.000004907086,0.000008829605,0.00012359465,0.000007505154,0.0000039460315,0.000020486772,0.00002716037,0.98961395,0.001132252,0.007952059,0.0010981079,0.0000072784915],"about_ca_topic_score_codex":0.010077187,"about_ca_topic_score_gemma":0.008350087,"teacher_disagreement_score":0.010077187,"about_ca_system_score_codex":0.0013562908,"about_ca_system_score_gemma":0.0008654885,"threshold_uncertainty_score":0.020037055},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.6266507062173049,"score_gpt":0.4639213162888603,"score_spread":0.1627293899284446,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3170000029","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005986065,0.00047862885,0.9697562,0.00019984097,0.00008409636,0.00016755625,0.0021299515,0.019246303,0.0019513443],"genre_scores_gemma":[0.07052478,0.0005248902,0.91810775,0.00019120284,0.000039016926,0.00039834552,0.0035884588,0.0039656498,0.0026599674],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991716,0.00019293395,0.000046365127,0.00017676018,0.00032921362,0.000083120205],"domain_scores_gemma":[0.9984365,0.0008341026,0.00007351385,0.00033717285,0.00021358841,0.00010495083],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009787785,0.0027122754,0.0017385242,0.0032166876,0.0009824823,0.0036195265,0.002749282,0.0016553695,0.015980171],"category_scores_gemma":[0.005527802,0.0012195946,0.002568166,0.0027266392,0.00076550816,0.004205901,0.008339535,0.0021437348,0.0035961117],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00087932235,0.00036778796,0.0040747956,0.0027102684,0.00047730395,0.00078184396,0.0046998067,0.045363747,0.040395353,0.05763375,0.05950159,0.7831145],"study_design_scores_gemma":[0.00018501887,0.0002918825,0.0012048464,0.00035179153,0.0001897784,0.0009878624,0.0018018247,0.74727595,0.030032968,0.10759352,0.10987735,0.00020723819],"about_ca_topic_score_codex":0.00359972,"about_ca_topic_score_gemma":0.012298469,"teacher_disagreement_score":0.015980171,"about_ca_system_score_codex":0.00048658598,"about_ca_system_score_gemma":0.0009897354,"threshold_uncertainty_score":0.05345905},"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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","score_opus":0.9041873969734017,"score_gpt":0.5915173642557057,"score_spread":0.312670032717696,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3208577093","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.35950917,0.0021201242,0.61339647,0.0062009236,0.00014964491,0.00018971431,0.0012286732,0.00042610767,0.016779054],"genre_scores_gemma":[0.95688194,0.0023493574,0.026962139,0.000282767,0.00021096709,0.00020985823,0.00096483965,0.000096548545,0.012041581],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9981687,0.0008956382,0.000087510816,0.00039404002,0.00028291438,0.00017115069],"domain_scores_gemma":[0.96624905,0.025328297,0.003907161,0.001731667,0.0017473138,0.001036485],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0032065941,0.0006868368,0.0011647935,0.0039047103,0.0017903355,0.0040052054,0.0016444048,0.0021316034,0.0055796267],"category_scores_gemma":[0.041441698,0.0009005244,0.0011581883,0.0032535582,0.0027029624,0.007529044,0.0029987535,0.0023393165,0.0007191086],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0001130074,0.000060808354,0.008531292,0.00016911874,0.00009829074,0.00028138206,0.0012505315,0.15938732,0.0007445651,0.80919117,0.004239142,0.015933366],"study_design_scores_gemma":[0.000027899461,0.000024151836,0.0013026602,0.000049289574,0.00003422407,0.00017822777,0.000343807,0.5519982,0.00025074484,0.4428458,0.0029237412,0.000021320546],"about_ca_topic_score_codex":0.0082272375,"about_ca_topic_score_gemma":0.0040029804,"teacher_disagreement_score":0.0082272375,"about_ca_system_score_codex":0.0020226173,"about_ca_system_score_gemma":0.0009783566,"threshold_uncertainty_score":0.018665671},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.2926327152332855,"score_gpt":0.3403536589814936,"score_spread":0.04772094374820812,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4295709232","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.67019516,0.0011780919,0.30842632,0.0020240077,0.00021276403,0.00033152723,0.0018969021,0.0061956435,0.009539519],"genre_scores_gemma":[0.8633912,0.00026382753,0.13028924,0.00029935004,0.000052033327,0.00028989383,0.0031484086,0.0005034576,0.0017624804],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.997387,0.0010543519,0.00011274163,0.0006078207,0.00064023054,0.00019781697],"domain_scores_gemma":[0.9777294,0.01395741,0.0012016278,0.003943409,0.0025330477,0.00063511357],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0062417295,0.0007606768,0.0010029682,0.0015449343,0.00091702986,0.0016203615,0.0021266877,0.0015441345,0.0016302072],"category_scores_gemma":[0.037279453,0.00052922766,0.0006605778,0.0011803962,0.0013045862,0.0031581705,0.0012615811,0.0015176067,0.0006041048],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0006084232,0.00030755432,0.0067680925,0.00020815068,0.00013141765,0.0001122484,0.00012380326,0.93315375,0.0040317844,0.017206358,0.006856188,0.030492289],"study_design_scores_gemma":[0.00003665467,0.00005061119,0.000272131,0.000005364736,0.0000066384077,0.000036074936,0.000012888492,0.99261224,0.0010415593,0.005539734,0.00037934456,0.0000066533216],"about_ca_topic_score_codex":0.008540486,"about_ca_topic_score_gemma":0.006863521,"teacher_disagreement_score":0.008540486,"about_ca_system_score_codex":0.0018190276,"about_ca_system_score_gemma":0.0014973497,"threshold_uncertainty_score":0.033009768},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.4518594465811662,"score_gpt":0.42245985450033313,"score_spread":0.029399592080833092,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405429927","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.019098341,0.00033165715,0.9771892,0.00022922884,0.00008278345,0.000060015933,0.00108266,0.0011520796,0.00077410054],"genre_scores_gemma":[0.4493339,0.0013312874,0.5368565,0.0001303415,0.0002641492,0.00054789754,0.0070795338,0.001102315,0.0033540968],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9977526,0.0010369588,0.0001323967,0.00033170512,0.00058603205,0.00016030486],"domain_scores_gemma":[0.9944165,0.0032600958,0.0003433363,0.0009685708,0.0008992756,0.00011233962],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002363251,0.0013035019,0.0008747523,0.0033020927,0.00057460024,0.0018736672,0.0006943041,0.00037895006,0.0021317485],"category_scores_gemma":[0.0113089755,0.00039219036,0.001610677,0.004391936,0.00075318734,0.0013428644,0.0010916861,0.0011916215,0.00088603114],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000296377,0.00021021548,0.013551572,0.0006139387,0.00081273186,0.00018605826,0.00046569316,0.13919058,0.011283219,0.083061084,0.01604392,0.7342846],"study_design_scores_gemma":[0.000020587184,0.0000527418,0.0143719865,0.00005817233,0.00012310574,0.00013536942,0.0001709007,0.88366777,0.0026551448,0.08946229,0.009224762,0.000057201945],"about_ca_topic_score_codex":0.0056733503,"about_ca_topic_score_gemma":0.003915562,"teacher_disagreement_score":0.0056733503,"about_ca_system_score_codex":0.0005238044,"about_ca_system_score_gemma":0.0017024623,"threshold_uncertainty_score":0.0124982},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.1808340169138115,"score_gpt":0.4031757345333891,"score_spread":0.2223417176195776,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413323615","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.049615566,0.00022306471,0.9463026,0.00044998984,0.00002034182,0.00007497349,0.00021761145,0.0021134384,0.000982393],"genre_scores_gemma":[0.53927594,0.00042293945,0.45536166,0.00021441207,0.00005661245,0.00016020847,0.0015413933,0.00034425803,0.002622621],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995511,0.0001232467,0.000015418658,0.00013195054,0.00013583031,0.00004244515],"domain_scores_gemma":[0.9985252,0.00086715433,0.00017772558,0.00022664995,0.0001696091,0.000033699398],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004633247,0.0010230123,0.00051177596,0.002356116,0.0004642303,0.0006415534,0.0008846417,0.00072404667,0.0013220375],"category_scores_gemma":[0.003259009,0.0003758807,0.0012761804,0.0008375891,0.0006980544,0.0013728942,0.0009646087,0.0012480083,0.0002998075],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000117402065,0.0002274563,0.0073544825,0.00018715371,0.00014884585,0.00040097669,0.00030540224,0.5718219,0.014275939,0.02755925,0.0041278126,0.3734734],"study_design_scores_gemma":[0.0000047011304,0.000017743234,0.00057451223,0.000007657289,0.000019896206,0.000048123373,0.000026146887,0.9753627,0.0020775176,0.020848172,0.0010069656,0.000005886414],"about_ca_topic_score_codex":0.005317615,"about_ca_topic_score_gemma":0.0095566595,"teacher_disagreement_score":0.005317615,"about_ca_system_score_codex":0.0008082842,"about_ca_system_score_gemma":0.00085521996,"threshold_uncertainty_score":0.010573328},"labels":[],"label_agreement":null}]}