{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":20,"total_is_capped":false,"direct_labels_cover":1,"predictions_cover":20,"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":"5d90b1ae8224","filters":{"venue":"International Journal of Data Mining and Bioinformatics"}},"results":[{"id":"W2072197604","doi":"10.1504/ijdmb.2011.038578","title":"SVM-RFE based feature selection for tandem mass spectrum quality assessment","year":2011,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":30,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Saskatchewan","funders":"University of Saskatchewan; Université Laval","keywords":"Computer science; Tandem; Support vector machine; Pattern recognition (psychology); Artificial intelligence; Feature selection; Quality (philosophy); Quality assessment; Feature (linguistics); Mass spectrum; Machine learning; Evaluation methods; Chemistry; Reliability engineering; Mass spectrometry; Physics; Engineering","authors":[{"name":"Jiarui Ding","is_ca":true},{"name":"Jinhong Shi","is_ca":true},{"name":"Fang‐Xiang Wu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0812831393055651,"gpt":0.3529058030219979,"spread":0.2716226637164327,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003869965,0.001125463,0.001712775,0.002187049,0.0003828147,0.0007127302,0.001129064,0.0009489756,0.001189954],"category_scores_gemma":[0.007043,0.0002335855,0.001198475,0.001378997,0.0002608663,0.0008265266,0.0005196797,0.0009767161,0.0007829675],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002973044,"about_ca_system_score_gemma":0.0005601064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00130886,"about_ca_topic_score_gemma":0.001180922,"domain_scores_codex":[0.998284,0.0005970905,0.0002013445,0.0002790357,0.000505943,0.0001325506],"domain_scores_gemma":[0.9964741,0.00178607,0.0003019491,0.0002240519,0.001157382,0.00005643866],"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":[0.0005633061,0.0004585393,0.008162682,0.0002947827,0.0002840282,0.0002489348,0.00006358231,0.04033889,0.03222221,0.0008083745,0.004709399,0.9118454],"study_design_scores_gemma":[0.0000747486,0.0003347947,0.00907948,0.00002715508,0.0001142201,0.0003077427,0.00004330228,0.96486,0.0219655,0.001331034,0.001816018,0.00004599427],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.05589991,0.0008726976,0.9395933,0.000188939,0.00005826837,0.000166986,0.0003802377,0.002415626,0.0004239145],"genre_scores_gemma":[0.5594476,0.0003017373,0.4367423,0.0001324742,0.00007717159,0.000349678,0.001891112,0.0001071934,0.0009506618],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003869965,"threshold_uncertainty_score":0.02046657,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2063036182","doi":"10.1504/ijdmb.2006.009920","title":"Bi-level clustering of mixed categorical and numerical biomedical data","year":2006,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":25,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Categorical variable; Cluster analysis; Data type; Computer science; Data mining; Artificial intelligence; Machine learning","authors":[{"name":"Bill Andreopoulos","is_ca":true},{"name":"Aijun An","is_ca":true},{"name":"Xiaogang Wang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1019047829298678,"gpt":0.3569539844642329,"spread":0.2550492015343652,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008788549,0.0008889947,0.001682444,0.006657368,0.001657125,0.004179042,0.002879651,0.001933671,0.001585164],"category_scores_gemma":[0.0387142,0.0007144228,0.002422013,0.007358713,0.001493843,0.002877817,0.004245415,0.002363413,0.001685008],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002069625,"about_ca_system_score_gemma":0.002620335,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003568818,"about_ca_topic_score_gemma":0.00529055,"domain_scores_codex":[0.9861137,0.004864834,0.001435435,0.002358521,0.004531794,0.0006956628],"domain_scores_gemma":[0.9808457,0.007942857,0.001509382,0.003903715,0.005318032,0.0004803087],"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":[0.001385105,0.0004713289,0.04300907,0.001376315,0.001063827,0.0004983288,0.002569375,0.1135197,0.04161668,0.07877936,0.01347529,0.7022356],"study_design_scores_gemma":[0.00008546578,0.0001684616,0.01270235,0.0001630075,0.000155921,0.0006013048,0.0005239307,0.7944673,0.01850422,0.157199,0.01524736,0.0001817246],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01417714,0.00027821,0.9830005,0.0002558116,0.00003959114,0.0001327103,0.0005236165,0.0009367215,0.0006557521],"genre_scores_gemma":[0.1135581,0.0001591445,0.8808381,0.0002541481,0.00004432452,0.000441393,0.00350873,0.0002048828,0.0009911795],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008788549,"threshold_uncertainty_score":0.04647887,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2067203347","doi":"10.1504/ijdmb.2008.016753","title":"Gene Regulatory Network modelling: a state-space approach","year":2008,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Gene Regulatory Network Analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Saskatchewan","funders":"Natural Sciences and Engineering Research Council of Canada; Dalian University of Technology","keywords":"Principal component analysis; Gene regulatory network; Probabilistic logic; State variable; Computer science; Bayesian network; State (computer science); State space; Control (management); Bayesian information criterion; Expression (computer science); Statistical model; State-space representation; Data mining; Gene; Mathematics; Artificial intelligence; Gene expression; Biology; Statistics; Genetics; Algorithm; Physics","authors":[{"name":"Fang‐Xiang Wu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0413704728274463,"gpt":0.2592375920972098,"spread":0.2178671192697635,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009978717,0.0009593382,0.001092029,0.001014899,0.0004618913,0.00134643,0.001737981,0.001185577,0.002027553],"category_scores_gemma":[0.003283791,0.0005302363,0.001316261,0.001128989,0.001272369,0.001786787,0.0008668565,0.001485082,0.0004379684],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001116684,"about_ca_system_score_gemma":0.001028077,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01226653,"about_ca_topic_score_gemma":0.007060498,"domain_scores_codex":[0.9993691,0.0002280982,0.00002664784,0.0002009311,0.0001286392,0.00004660733],"domain_scores_gemma":[0.9987487,0.0009037965,0.0001347314,0.00006763241,0.0001119984,0.0000330811],"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":[0.00002290737,0.00001732034,0.0005556197,0.00002828335,0.00003200934,0.00003871372,0.00004714882,0.9610862,0.0005577023,0.03018732,0.0001863244,0.007240385],"study_design_scores_gemma":[0.000001890188,0.0000045485,0.0000414123,0.000002084241,0.000005183905,0.000004209239,0.000002327575,0.9924967,0.00007440796,0.007197274,0.0001668653,0.00000300823],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006926893,0.0001933765,0.9913408,0.000148081,0.00001532898,0.00001732052,0.0001100939,0.000225846,0.001022229],"genre_scores_gemma":[0.7866635,0.00129327,0.2049969,0.0001497367,0.0001283173,0.0005001894,0.0007234822,0.00009835103,0.005446253],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01226653,"threshold_uncertainty_score":0.02439022,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4252436249","doi":"10.1504/ijdmb.2018.094776","title":"Integration of k-means clustering algorithm with network analysis for drug-target interactions network prediction","year":2018,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Cluster analysis; Bayes' theorem; Naive Bayes classifier; Computer science; Artificial intelligence; Drug discovery; Machine learning; Interaction network; In silico; Support vector machine; Classifier (UML); Computational biology; Pattern recognition (psychology); Data mining; Bioinformatics; Biology; Bayesian probability","authors":[{"name":"Sara Aghakhani","is_ca":true},{"name":"Ala Qabaja","is_ca":true},{"name":"Reda Alhajj","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04307975514025782,"gpt":0.3345422988701792,"spread":0.2914625437299214,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009986616,0.0009407758,0.001033997,0.004264479,0.0008541076,0.0006752654,0.0007665407,0.000822859,0.001072476],"category_scores_gemma":[0.00263927,0.0004062943,0.001092985,0.002167973,0.0003319865,0.001224095,0.0005936473,0.0007686265,0.000479935],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000945556,"about_ca_system_score_gemma":0.001533343,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01476479,"about_ca_topic_score_gemma":0.01518761,"domain_scores_codex":[0.9992278,0.0001818154,0.00006654881,0.0001694974,0.0002928112,0.00006157882],"domain_scores_gemma":[0.9990051,0.0004625583,0.0001356668,0.00006509024,0.0002925037,0.00003914432],"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":[0.0002634414,0.0002889772,0.01452361,0.0003901147,0.0005342776,0.0001963493,0.0002147754,0.4946794,0.007439236,0.008249397,0.005789866,0.4674306],"study_design_scores_gemma":[0.00000962624,0.00002721162,0.001342364,0.00001153527,0.00002633735,0.00005237429,0.00002553612,0.9914414,0.001500131,0.004489638,0.001055513,0.0000184427],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02572313,0.0006668029,0.9694735,0.0001918743,0.00005020312,0.0001721241,0.0003708882,0.001798168,0.001553312],"genre_scores_gemma":[0.3005241,0.0006594955,0.6957562,0.00007381535,0.00005821715,0.0003488411,0.001086219,0.0001155943,0.00137756],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01476479,"threshold_uncertainty_score":0.02935767,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1975341743","doi":"10.1504/ijdmb.2010.035898","title":"Using Gene Ontology to enhance effectiveness of similarity measures for microarray data","year":2010,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Redundancy (engineering); Gene ontology; Computer science; Semantic similarity; Data mining; Feature selection; Similarity (geometry); Gene selection; Microarray analysis techniques; Artificial intelligence; Pattern recognition (psychology); Gene; Gene expression; Biology; Genetics","authors":[{"name":"Zheng Chen","is_ca":true},{"name":"Jian Tang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08225493591224538,"gpt":0.400972716185559,"spread":0.3187177802733136,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004495233,0.0006412428,0.001411598,0.004881209,0.0005268084,0.00133446,0.0009923091,0.0006803724,0.0005142554],"category_scores_gemma":[0.01561788,0.0002437252,0.001077497,0.005055957,0.0006337696,0.002842846,0.001419049,0.0008681977,0.0002975724],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000658806,"about_ca_system_score_gemma":0.0006666852,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006492126,"about_ca_topic_score_gemma":0.0008487502,"domain_scores_codex":[0.9951419,0.001475013,0.0005415822,0.0005130348,0.002157782,0.0001707748],"domain_scores_gemma":[0.992391,0.00456318,0.0007866278,0.001191213,0.0009433257,0.0001246355],"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":[0.000491652,0.0003994271,0.01588321,0.0008411892,0.0005792248,0.0003401512,0.0004970448,0.04937797,0.0969817,0.01948273,0.002651171,0.8124745],"study_design_scores_gemma":[0.0001033821,0.0006780485,0.01997127,0.0001015606,0.0002350971,0.001350677,0.0003312056,0.8168432,0.07757825,0.07335858,0.009302979,0.0001457289],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07370114,0.0007602567,0.9231595,0.0001860204,0.00006072383,0.0001270827,0.0002333608,0.0008847952,0.0008871445],"genre_scores_gemma":[0.5008396,0.0005018849,0.4967857,0.0001170121,0.0001044318,0.0002556824,0.0008390027,0.0001269693,0.0004297658],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004881209,"threshold_uncertainty_score":0.02377337,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2018695726","doi":"10.1504/ijdmb.2011.040384","title":"Image segmentation of biofilm structures using optimal multi-level thresholding","year":2011,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Bacterial biofilms and quorum sensing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":7,"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":"Universidad de Atacama; Ontario Innovation Trust","keywords":"Thresholding; Segmentation; Artificial intelligence; Image segmentation; Computer science; Cluster analysis; Pattern recognition (psychology); Rand index; Process (computing); Set (abstract data type); Computer vision; Image (mathematics)","authors":[{"name":"Darío Rojas","is_ca":false},{"name":"Luis Rueda","is_ca":true},{"name":"Alioune Ngom","is_ca":true},{"name":"Homero Hurrutia","is_ca":false},{"name":"Gerardo Cárcamo‐Oyarce","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1163285617521467,"gpt":0.3320784984421196,"spread":0.2157499366899729,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00206439,0.0005868831,0.00114083,0.00280093,0.0006852791,0.001833943,0.001069108,0.00137919,0.001041228],"category_scores_gemma":[0.004574033,0.0006709107,0.001295737,0.001647153,0.0009560228,0.001402997,0.0009927124,0.001084521,0.0005679363],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00085125,"about_ca_system_score_gemma":0.001151699,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001626301,"about_ca_topic_score_gemma":0.001905045,"domain_scores_codex":[0.998447,0.0002945871,0.000162239,0.0003652602,0.000578143,0.0001529089],"domain_scores_gemma":[0.9981458,0.0007373497,0.0002718657,0.0002513083,0.0005301089,0.00006355342],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000360749,0.0001457785,0.002541117,0.0006478387,0.0001263273,0.0002273878,0.0006410102,0.0549766,0.5786582,0.005417928,0.001165613,0.3550915],"study_design_scores_gemma":[0.00003218274,0.0002612844,0.007581606,0.00009802885,0.0001063287,0.0005810446,0.0002336945,0.6403051,0.3369967,0.009231878,0.00443806,0.0001340413],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02752047,0.0002961447,0.9706904,0.0001062102,0.00002307133,0.00008900208,0.00004965418,0.000612796,0.0006122623],"genre_scores_gemma":[0.1181919,0.0002439381,0.8808261,0.00003831093,0.00001051683,0.00008472759,0.0001101849,0.0001417476,0.0003525809],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00280093,"threshold_uncertainty_score":0.01091766,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2146392361","doi":"10.1504/ijdmb.2013.056616","title":"A robust hidden semi-Markov model with application to aCGH data processing","year":2013,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Gene expression and cancer classification","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 British Columbia; BC Cancer Agency; University of British Columbia Hospital","funders":"University of British Columbia","keywords":"Outlier; Hidden Markov model; Markov model; Computer science; Benchmark (surveying); Glioblastoma; Comparative genomic hybridization; Markov chain; Reliability (semiconductor); Data-driven; Data mining; Artificial intelligence; Machine learning; Biology; Power (physics)","authors":[{"name":"Jiarui Ding","is_ca":true},{"name":"Sohrab P. Shah","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04764101485050556,"gpt":0.3031798806265017,"spread":0.2555388657759962,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002500209,0.0008037197,0.001090547,0.0006770665,0.0005610895,0.001007357,0.001655009,0.001202301,0.001938003],"category_scores_gemma":[0.00743671,0.0008511433,0.001246315,0.0008306211,0.0008423721,0.001179741,0.001160529,0.002267398,0.001114522],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009744067,"about_ca_system_score_gemma":0.001734735,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01207188,"about_ca_topic_score_gemma":0.01031653,"domain_scores_codex":[0.9989831,0.0003768871,0.00006922643,0.000261835,0.0002226072,0.00008640629],"domain_scores_gemma":[0.9960703,0.003025652,0.0002845623,0.0002121853,0.0003382321,0.00006926606],"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":[0.0001835532,0.0000494417,0.0008716657,0.00007635963,0.0000919488,0.0001060167,0.00009417582,0.9089769,0.004177429,0.01217631,0.0009127759,0.07228346],"study_design_scores_gemma":[0.000004085387,0.00001016503,0.00005198423,0.000002552764,0.000004922704,0.00001331016,0.000002556548,0.9962025,0.0005021326,0.003003531,0.0001936432,0.000008645838],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002072335,0.00006187877,0.9972934,0.00006100467,0.00000985096,0.000014669,0.00004560902,0.0003530351,0.00008807454],"genre_scores_gemma":[0.3103127,0.0005070503,0.684301,0.0002026902,0.0001105367,0.0004689277,0.0009036774,0.000278966,0.002914365],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01207188,"threshold_uncertainty_score":0.02400321,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2891379954","doi":"10.1504/ijdmb.2018.10016075","title":"Integration of k-means clustering algorithm with network analysis for drug-target interactions network prediction","year":2018,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Cluster analysis; Computer science; Drug; Algorithm; Network analysis; k-means clustering; Data mining; Artificial intelligence; Biology; Pharmacology; Physics","authors":[{"name":"Sara Aghakhani","is_ca":true},{"name":"Ala Qabaja","is_ca":true},{"name":"Reda Alhajj","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04307975514025782,"gpt":0.3345422988701792,"spread":0.2914625437299214,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009986616,0.0009407758,0.001033997,0.004264479,0.0008541076,0.0006752654,0.0007665407,0.000822859,0.001072476],"category_scores_gemma":[0.00263927,0.0004062943,0.001092985,0.002167973,0.0003319865,0.001224095,0.0005936473,0.0007686265,0.000479935],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000945556,"about_ca_system_score_gemma":0.001533343,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01476479,"about_ca_topic_score_gemma":0.01518761,"domain_scores_codex":[0.9992278,0.0001818154,0.00006654881,0.0001694974,0.0002928112,0.00006157882],"domain_scores_gemma":[0.9990051,0.0004625583,0.0001356668,0.00006509024,0.0002925037,0.00003914432],"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":[0.0002634414,0.0002889772,0.01452361,0.0003901147,0.0005342776,0.0001963493,0.0002147754,0.4946794,0.007439236,0.008249397,0.005789866,0.4674306],"study_design_scores_gemma":[0.00000962624,0.00002721162,0.001342364,0.00001153527,0.00002633735,0.00005237429,0.00002553612,0.9914414,0.001500131,0.004489638,0.001055513,0.0000184427],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02572313,0.0006668029,0.9694735,0.0001918743,0.00005020312,0.0001721241,0.0003708882,0.001798168,0.001553312],"genre_scores_gemma":[0.3005241,0.0006594955,0.6957562,0.00007381535,0.00005821715,0.0003488411,0.001086219,0.0001155943,0.00137756],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01476479,"threshold_uncertainty_score":0.02935767,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2119632600","doi":"10.1504/ijdmb.2011.040388","title":"Combining multiple perspective as intelligent agents into robust approach for biomarker detection in gene expression data","year":2011,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Evolutionary Algorithms and Applications","field":"Computer Science","cited_by":3,"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":"Perspective (graphical); Gene; Gene expression; Biomarker; Computer science; Computational biology; Expression (computer science); microRNA; Data mining; Biology; Artificial intelligence; Genetics","authors":[{"name":"Mohammed Alshalalfa","is_ca":true},{"name":"Ghada Naji","is_ca":false},{"name":"Ala Qabaja","is_ca":true},{"name":"Reda Alhajj","is_ca":true},{"name":"Jon Rokne","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1596503619878044,"gpt":0.336622064281259,"spread":0.1769717022934547,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003672507,0.001180944,0.001421329,0.001292674,0.00053805,0.002026777,0.00148793,0.001215386,0.0009410146],"category_scores_gemma":[0.006300414,0.0005310532,0.001427984,0.001060332,0.0009304212,0.002002161,0.001903092,0.001136768,0.0002177673],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008634284,"about_ca_system_score_gemma":0.0009445988,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001365743,"about_ca_topic_score_gemma":0.001874154,"domain_scores_codex":[0.9974299,0.0010805,0.0001529448,0.0004771379,0.0007180125,0.000141427],"domain_scores_gemma":[0.9972101,0.001521037,0.0004762186,0.0002677119,0.0003981022,0.0001268091],"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":[0.0004165865,0.0002225967,0.008367579,0.0003714139,0.0007628634,0.0005879054,0.0004754433,0.6791292,0.01897564,0.05124645,0.0009327946,0.2385114],"study_design_scores_gemma":[0.00001650306,0.0001007148,0.0004702723,0.00001019328,0.00008672701,0.00006332286,0.00003930089,0.9860581,0.002313882,0.009810519,0.001013823,0.00001660596],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01773013,0.0003392296,0.9807137,0.0001952231,0.00002127578,0.00003980013,0.00002283695,0.0001584701,0.000779461],"genre_scores_gemma":[0.4481933,0.0003613713,0.5496056,0.0001474534,0.00006474879,0.0001319419,0.0001179902,0.00004503684,0.001332631],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003672507,"threshold_uncertainty_score":0.01942235,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2150689251","doi":"10.1504/ijdmb.2013.054696","title":"A clustering approach for estimating parameters of a profile hidden Markov model","year":2013,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Bayesian Methods and Mixture Models","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":"Université de Montréal","funders":"University of Waterloo; University of Tehran; Iran National Science Foundation","keywords":"Hidden Markov model; Cluster analysis; Pattern recognition (psychology); Computer science; Markov chain Monte Carlo; Markov chain; Artificial intelligence; Markov model; Bayesian probability; Mathematics; Algorithm; Machine learning","authors":[{"name":"Rosa Aghdam","is_ca":false},{"name":"Hamid Pezeshk","is_ca":false},{"name":"Seyed Amir Malekpour","is_ca":true},{"name":"Soudabeh Shemehsavar","is_ca":false},{"name":"Changiz Eslahchi","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06555281677021664,"gpt":0.3188018336758385,"spread":0.2532490169056219,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003321627,0.001228181,0.001347383,0.003333425,0.001427648,0.001181924,0.002917329,0.001904683,0.001695862],"category_scores_gemma":[0.01127967,0.00101454,0.002093515,0.002852423,0.0006419956,0.001975889,0.001298168,0.002680098,0.001525286],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001242616,"about_ca_system_score_gemma":0.002421128,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01185177,"about_ca_topic_score_gemma":0.01296537,"domain_scores_codex":[0.9975535,0.0007873822,0.0001754834,0.0007056532,0.000653661,0.0001242285],"domain_scores_gemma":[0.9965099,0.001724249,0.0002337657,0.0005471893,0.0008835211,0.0001014011],"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":[0.0002717074,0.0002601013,0.006979713,0.0003358388,0.0005362267,0.0002233042,0.0004994543,0.3092438,0.01182931,0.03782035,0.005165458,0.6268347],"study_design_scores_gemma":[0.00001517308,0.00003238681,0.0008909398,0.00003014022,0.00005179515,0.0001284608,0.00003635919,0.9737394,0.002939749,0.0197087,0.002369766,0.00005710664],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001813691,0.00008508334,0.9974848,0.00003020198,0.00001234527,0.00004312311,0.00006433063,0.0003349643,0.0001314032],"genre_scores_gemma":[0.05372161,0.0002265396,0.9440271,0.00006974034,0.00003522118,0.0002582726,0.0008405134,0.0001390382,0.0006820005],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01185177,"threshold_uncertainty_score":0.02356559,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":[],"domain":null,"study_design":"simulation_or_modeling","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4230716249","doi":"10.1504/ijdmb.2020.112850","title":"MHCherryPan: a novel pan-specific model for binding affinity prediction of class I HLA-peptide","year":2020,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"vaccines and immunoinformatics approaches","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Human leukocyte antigen; Benchmark (surveying); Computational biology; Allele; Immune system; Artificial immune system; Biology; Computer science; Class (philosophy); Peptide; Artificial intelligence; Antigen; Immunology; Gene; Genetics; Biochemistry; Geography","authors":[{"name":"Xuezhi Xie","is_ca":true},{"name":"Yuanyuan Han","is_ca":true},{"name":"Kaizhong Zhang","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1157585872825637,"gpt":0.2929261544225714,"spread":0.1771675671400076,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004331772,0.0006790648,0.0006472626,0.0005445485,0.0002674568,0.0004583854,0.001625612,0.0008081337,0.001328482],"category_scores_gemma":[0.0005883135,0.0002991602,0.0005981874,0.0004432229,0.0002380712,0.0007361844,0.0005937798,0.0008871986,0.0003171219],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007413374,"about_ca_system_score_gemma":0.001079196,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0168054,"about_ca_topic_score_gemma":0.01598828,"domain_scores_codex":[0.9998813,0.00001809574,0.00000674096,0.0000404556,0.00002723977,0.00002607172],"domain_scores_gemma":[0.9998429,0.00005462466,0.00001877811,0.00001308987,0.00005436662,0.00001619459],"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":[0.0002093225,0.0001648775,0.004002255,0.00004882117,0.0001314028,0.0001423183,0.00002521804,0.8672113,0.006898559,0.00273406,0.002569416,0.1158625],"study_design_scores_gemma":[0.000002633253,0.00000747546,0.00008504629,8.739561e-7,0.000003810969,0.000006339931,7.000461e-7,0.9992656,0.000257105,0.0002529496,0.0001164061,0.000001200079],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1892129,0.001354686,0.8012974,0.0006990182,0.0001609397,0.0001139483,0.0008234044,0.002955126,0.003382547],"genre_scores_gemma":[0.8508148,0.0006551487,0.1376966,0.0004512263,0.00008995576,0.0001885812,0.001571305,0.0001214763,0.008410945],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0168054,"threshold_uncertainty_score":0.0334152,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2045369746","doi":"10.1504/ijdmb.2013.056090","title":"Integrating machine learning techniques into robust data enrichment approach and its application to gene expression data","year":2013,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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":"Computer science; Machine learning; Sample (material); Data mining; Process (computing); Artificial intelligence; Domain (mathematical analysis); Expression (computer science); Representation (politics); Independence (probability theory); Domain knowledge; Probabilistic logic","authors":[{"name":"Utku Erdoğdu","is_ca":false},{"name":"Mehmet Tan","is_ca":false},{"name":"Reda Alhajj","is_ca":true},{"name":"Faruk Polat","is_ca":false},{"name":"Jon Rokne","is_ca":true},{"name":"Douglas J. Demetrick","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04640594359035279,"gpt":0.3157167867357922,"spread":0.2693108431454394,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005214515,0.0009775767,0.001265098,0.002668941,0.0005627051,0.001145794,0.001343105,0.001123429,0.0008134216],"category_scores_gemma":[0.008970113,0.0004776895,0.002238238,0.002260945,0.0008955534,0.001109895,0.001626185,0.001482269,0.0006164247],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006274102,"about_ca_system_score_gemma":0.001361892,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0025366,"about_ca_topic_score_gemma":0.002486102,"domain_scores_codex":[0.9959681,0.001551783,0.0002289904,0.000891448,0.001193221,0.0001664827],"domain_scores_gemma":[0.9958182,0.002429883,0.0002637177,0.0007605391,0.0006524499,0.0000752017],"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":[0.0002304422,0.000288061,0.005595671,0.0004279789,0.0003922513,0.0005173189,0.0003040446,0.2568981,0.04674844,0.01458054,0.001526684,0.6724904],"study_design_scores_gemma":[0.00001388344,0.0001183772,0.001097262,0.00002496137,0.00006785721,0.0002145584,0.00005368889,0.9516526,0.02757451,0.01498715,0.004156576,0.00003860061],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.004886871,0.0001422099,0.9936372,0.00007450242,0.000009049665,0.00004079247,0.00004335215,0.001029464,0.0001365653],"genre_scores_gemma":[0.08462568,0.0001620786,0.9138613,0.0001209379,0.00002081992,0.0001336233,0.0004018459,0.0001569438,0.0005167915],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005214515,"threshold_uncertainty_score":0.02757728,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2946233355","doi":"10.1504/ijdmb.2019.099714","title":"Effective induction of gene regulatory networks using a novel recommendation method","year":2019,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Gene Regulatory Network Analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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":"Türkiye Bilimsel ve Teknolojik Araştırma Kurumu","keywords":"Computer science; Data mining; Inference; Set (abstract data type); Receiver operating characteristic; Gene regulatory network; Data set; In silico; Precision and recall; Measure (data warehouse); Collaborative filtering; Machine learning; Artificial intelligence; Gene; Recommender system; Biology; Genetics","authors":[{"name":"Makbule Gülçin Özsoy","is_ca":false},{"name":"Faruk Polat","is_ca":false},{"name":"Reda Alhajj","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02846873594889496,"gpt":0.3166203865312484,"spread":0.2881516505823534,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001155492,0.0008720576,0.001390726,0.002753482,0.0006475269,0.0008743356,0.002025953,0.001564211,0.002436496],"category_scores_gemma":[0.003910817,0.000539245,0.001589679,0.001895345,0.0004477271,0.001188593,0.0005570462,0.001203247,0.001352288],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007970265,"about_ca_system_score_gemma":0.001108704,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01072119,"about_ca_topic_score_gemma":0.01468289,"domain_scores_codex":[0.9985948,0.0002237024,0.00009180169,0.0005419888,0.000453618,0.00009402604],"domain_scores_gemma":[0.9980598,0.0008899919,0.0001686538,0.0003414365,0.0004747275,0.00006527755],"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":[0.0002398249,0.0003451513,0.00726072,0.0002952284,0.0003207056,0.0003502806,0.0001950648,0.1759787,0.02779475,0.01083142,0.006459001,0.7699291],"study_design_scores_gemma":[0.00003252742,0.00004843036,0.0008918255,0.00002073676,0.00005420246,0.0001719254,0.00001956006,0.9868608,0.005133918,0.004363815,0.002377491,0.00002483544],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005605408,0.0002138253,0.9919335,0.0001124705,0.00002722154,0.00006298687,0.0001695119,0.001111331,0.0007637889],"genre_scores_gemma":[0.1348141,0.0003654365,0.8588982,0.0002033641,0.0001353079,0.0002976612,0.001211083,0.0001488212,0.003926006],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01072119,"threshold_uncertainty_score":0.02131754,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3063921741","doi":"10.1504/ijdmb.2020.10031424","title":"Identification of protein hot regions by combing structure-based classification, energy-based clustering and sequence-based conservation in evolution","year":2020,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Protein Structure and Dynamics","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":"Simon Fraser University","funders":"","keywords":"Combing; Cluster analysis; Computer science; Sequence (biology); Identification (biology); Data mining; Hot spot (computer programming); Pattern recognition (psychology); Biological system; Artificial intelligence; Materials science; Biology; Genetics","authors":[{"name":"Nansheng Chen","is_ca":true},{"name":"Xiaolong Zhang","is_ca":false},{"name":"Haomin Gan","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03435348076170269,"gpt":0.2755679535106321,"spread":0.2412144727489294,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005034803,0.0004233657,0.0006303606,0.002561808,0.0005900518,0.0005508239,0.0006546336,0.0005388416,0.0005545312],"category_scores_gemma":[0.0009780444,0.0002380425,0.0006200738,0.001694741,0.0003375927,0.0007999765,0.0004522583,0.0003664259,0.0002569773],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003594742,"about_ca_system_score_gemma":0.0004238314,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003093842,"about_ca_topic_score_gemma":0.004077433,"domain_scores_codex":[0.9995666,0.00006080929,0.0000385555,0.000158621,0.0001248835,0.00005048243],"domain_scores_gemma":[0.9994785,0.0001342695,0.0001274554,0.00004876543,0.0001642412,0.00004684362],"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":[0.0007393442,0.0004745171,0.1698533,0.0005414227,0.0004985918,0.0006455273,0.0006973936,0.07214291,0.1860029,0.004991844,0.003177809,0.5602345],"study_design_scores_gemma":[0.00003696337,0.0001861579,0.08734952,0.00004868107,0.0001752189,0.0008958804,0.0002831207,0.8604198,0.0423507,0.005475211,0.002700527,0.00007833522],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6681511,0.001527966,0.3266059,0.0001925805,0.00004033063,0.0001458099,0.0003744127,0.0010794,0.001882553],"genre_scores_gemma":[0.8654767,0.0003279333,0.1320573,0.00005892667,0.0000202693,0.00006761984,0.0008052075,0.00007812653,0.001107986],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003093842,"threshold_uncertainty_score":0.006151676,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2103153231","doi":"10.1504/ijdmb.2012.048172","title":"Re-ranking with context for high-performance biomedical information retrieval","year":2012,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Biomedical Text Mining and Ontologies","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":"York University","funders":"U.S. National Library of Medicine","keywords":"Ranking (information retrieval); Context (archaeology); Information retrieval; Computer science; Data science; Biology; Paleontology","authors":[{"name":"Xiaoshi Yin","is_ca":false},{"name":"Jimmy Xiangji Huang","is_ca":true},{"name":"Zhoujun Li","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0339923099069285,"gpt":0.3003697377014244,"spread":0.266377427794496,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003901913,0.001654339,0.002469364,0.007882127,0.001357228,0.002383003,0.001938105,0.001255426,0.002483819],"category_scores_gemma":[0.01952,0.0005965707,0.001101447,0.004752973,0.0005498593,0.003059414,0.001648996,0.001448547,0.002258942],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008095561,"about_ca_system_score_gemma":0.001605781,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003595286,"about_ca_topic_score_gemma":0.009043274,"domain_scores_codex":[0.9935336,0.002840369,0.0004550014,0.0007245449,0.002157319,0.0002890764],"domain_scores_gemma":[0.9879206,0.005457501,0.0007925127,0.002077043,0.00345053,0.0003017814],"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":[0.0008856748,0.0006273304,0.007416855,0.0009610887,0.0003923329,0.0006184091,0.0004072895,0.04025604,0.04815581,0.005880094,0.01476471,0.8796344],"study_design_scores_gemma":[0.0003412181,0.001643668,0.01418014,0.0002160488,0.0006794527,0.002281799,0.0005956879,0.8318621,0.07852088,0.04085069,0.02835011,0.0004782358],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09663832,0.01317451,0.8740811,0.0009583925,0.0005673135,0.0009272879,0.001106903,0.008806895,0.003739275],"genre_scores_gemma":[0.4441288,0.001930663,0.5477408,0.0003596257,0.0006506378,0.0003585251,0.002315557,0.0003080148,0.00220736],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007882127,"threshold_uncertainty_score":0.02063555,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2130013930","doi":"10.1504/ijdmb.2013.056080","title":"Subtree selection in kernels for graph classification","year":2013,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Computational Drug Discovery Methods","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 Calgary","funders":"","keywords":"Cheminformatics; Feature selection; Computer science; In silico; Graph; Selection (genetic algorithm); Feature (linguistics); Data mining; Artificial intelligence; Machine learning; Pattern recognition (psychology); Theoretical computer science; Bioinformatics; Biology","authors":[{"name":"Mehmet Tan","is_ca":false},{"name":"Faruk Polat","is_ca":false},{"name":"Reda Alhajj","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09114874666879506,"gpt":0.3597068872948939,"spread":0.2685581406260988,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001011122,0.0004472476,0.0008627273,0.001550287,0.0003705683,0.0006533636,0.0008058471,0.000581492,0.001014341],"category_scores_gemma":[0.004744747,0.0001986398,0.0006952924,0.001691986,0.0005035076,0.001131125,0.0007596402,0.0007604223,0.0004820015],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004561443,"about_ca_system_score_gemma":0.0005192338,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001747015,"about_ca_topic_score_gemma":0.001989427,"domain_scores_codex":[0.9989821,0.0003450307,0.0000769089,0.0002054083,0.0003133253,0.00007711013],"domain_scores_gemma":[0.9970551,0.001679463,0.000240278,0.0004507902,0.0004619335,0.0001124377],"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":[0.0003814993,0.0002651386,0.006405387,0.0001699596,0.0001088507,0.0002108648,0.0001596948,0.1422184,0.03249479,0.02706492,0.004309974,0.7862105],"study_design_scores_gemma":[0.00001250542,0.00006038974,0.00117551,0.000004968413,0.0000136743,0.00009544084,0.00002054236,0.9775516,0.003686956,0.01629074,0.001075662,0.00001201993],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04809575,0.0001952257,0.9501615,0.0000936255,0.00002040262,0.00004000044,0.000125114,0.0009636848,0.0003047062],"genre_scores_gemma":[0.5519561,0.0002144146,0.4449005,0.00006795198,0.00006377327,0.0001505276,0.001051738,0.0001719263,0.001423071],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001747015,"threshold_uncertainty_score":0.005347371,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3137172717","doi":"10.1504/ijdmb.2020.10036329","title":"A network enhancement-based method for clustering of single cell RNA-seq data","year":2020,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Single-cell and spatial transcriptomics","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":"University of Saskatchewan","funders":"","keywords":"Cluster analysis; Computer science; Similarity (geometry); Data mining; Node (physics); Noise (video); Artificial intelligence; Pattern recognition (psychology); Physics","authors":[{"name":"Hong Dong Li","is_ca":false},{"name":"Fang‐Xiang Wu","is_ca":true},{"name":"Xiaoqing Peng","is_ca":false},{"name":"Yunpei Xu","is_ca":false},{"name":"Rongyuan Li","is_ca":false},{"name":"Xiaoshu Zhu","is_ca":false},{"name":"Lilu Guo","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09629500041666537,"gpt":0.3209526518611276,"spread":0.2246576514444623,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001263484,0.00106408,0.0008396587,0.002768983,0.0009672371,0.0007633076,0.001467242,0.001011899,0.001256734],"category_scores_gemma":[0.00343569,0.000505753,0.001418069,0.002422906,0.0006620602,0.001344481,0.001479808,0.001219582,0.000874789],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009033934,"about_ca_system_score_gemma":0.001282003,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005491725,"about_ca_topic_score_gemma":0.009944635,"domain_scores_codex":[0.998583,0.0002255945,0.00007556137,0.0005277251,0.0004802763,0.0001077869],"domain_scores_gemma":[0.9986579,0.0003630771,0.0001824819,0.0002079612,0.0004907191,0.00009774646],"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":[0.0005775696,0.0002439717,0.005564078,0.0005338705,0.0004020885,0.0005270707,0.0006646224,0.1258477,0.1470859,0.008711333,0.007223864,0.7026179],"study_design_scores_gemma":[0.00003505506,0.00008561352,0.003089088,0.00001781873,0.00005535363,0.0003410387,0.0001045059,0.9468786,0.03554242,0.006234701,0.007548381,0.00006737581],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007040991,0.0001127128,0.9911727,0.00005612899,0.00002928554,0.00008036318,0.0001186937,0.001094355,0.0002946839],"genre_scores_gemma":[0.08291238,0.0001818212,0.9132122,0.00008925936,0.00004547255,0.0002769545,0.001087246,0.0002345789,0.001960146],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005491725,"threshold_uncertainty_score":0.01091951,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3214325870","doi":"10.1504/ijdmb.2020.113690","title":"A network enhancement-based method for clustering of single cell RNA-seq data","year":2020,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Single-cell and spatial transcriptomics","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":"University of Saskatchewan","funders":"","keywords":"Cluster analysis; Computer science; Similarity (geometry); Data mining; Node (physics); Noise (video); Artificial intelligence; Pattern recognition (psychology)","authors":[{"name":"Xiaoshu Zhu","is_ca":false},{"name":"Lilu Guo","is_ca":false},{"name":"Rongyuan Li","is_ca":false},{"name":"Yunpei Xu","is_ca":false},{"name":"Fang‐Xiang Wu","is_ca":true},{"name":"Xiaoqing Peng","is_ca":false},{"name":"Hong Dong Li","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09629500041666537,"gpt":0.3209526518611276,"spread":0.2246576514444623,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001263484,0.00106408,0.0008396587,0.002768983,0.0009672371,0.0007633076,0.001467242,0.001011899,0.001256734],"category_scores_gemma":[0.00343569,0.000505753,0.001418069,0.002422906,0.0006620602,0.001344481,0.001479808,0.001219582,0.000874789],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009033934,"about_ca_system_score_gemma":0.001282003,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005491725,"about_ca_topic_score_gemma":0.009944635,"domain_scores_codex":[0.998583,0.0002255945,0.00007556137,0.0005277251,0.0004802763,0.0001077869],"domain_scores_gemma":[0.9986579,0.0003630771,0.0001824819,0.0002079612,0.0004907191,0.00009774646],"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":[0.0005775696,0.0002439717,0.005564078,0.0005338705,0.0004020885,0.0005270707,0.0006646224,0.1258477,0.1470859,0.008711333,0.007223864,0.7026179],"study_design_scores_gemma":[0.00003505506,0.00008561352,0.003089088,0.00001781873,0.00005535363,0.0003410387,0.0001045059,0.9468786,0.03554242,0.006234701,0.007548381,0.00006737581],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007040991,0.0001127128,0.9911727,0.00005612899,0.00002928554,0.00008036318,0.0001186937,0.001094355,0.0002946839],"genre_scores_gemma":[0.08291238,0.0001818212,0.9132122,0.00008925936,0.00004547255,0.0002769545,0.001087246,0.0002345789,0.001960146],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005491725,"threshold_uncertainty_score":0.01091951,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3107387809","doi":"10.1504/ijdmb.2020.10033716","title":"MHCherryPan: a novel pan-specific model for binding affinity prediction of class I HLA-peptide","year":2020,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"vaccines and immunoinformatics approaches","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":"Western University","funders":"","keywords":"Human leukocyte antigen; Benchmark (surveying); Computational biology; Allele; Immune system; Artificial immune system; Peptide; Computer science; Biology; Class (philosophy); Artificial intelligence; Antigen; Immunology; Gene; Genetics; Biochemistry; Geography","authors":[{"name":"Kaizhong Zhang","is_ca":true},{"name":"Yuanyuan Han","is_ca":true},{"name":"Xuezhi Xie","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1157585872825637,"gpt":0.2929261544225714,"spread":0.1771675671400076,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004331772,0.0006790648,0.0006472626,0.0005445485,0.0002674568,0.0004583854,0.001625612,0.0008081337,0.001328482],"category_scores_gemma":[0.0005883135,0.0002991602,0.0005981874,0.0004432229,0.0002380712,0.0007361844,0.0005937798,0.0008871986,0.0003171219],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007413374,"about_ca_system_score_gemma":0.001079196,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0168054,"about_ca_topic_score_gemma":0.01598828,"domain_scores_codex":[0.9998813,0.00001809574,0.00000674096,0.0000404556,0.00002723977,0.00002607172],"domain_scores_gemma":[0.9998429,0.00005462466,0.00001877811,0.00001308987,0.00005436662,0.00001619459],"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":[0.0002093225,0.0001648775,0.004002255,0.00004882117,0.0001314028,0.0001423183,0.00002521804,0.8672113,0.006898559,0.00273406,0.002569416,0.1158625],"study_design_scores_gemma":[0.000002633253,0.00000747546,0.00008504629,8.739561e-7,0.000003810969,0.000006339931,7.000461e-7,0.9992656,0.000257105,0.0002529496,0.0001164061,0.000001200079],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1892129,0.001354686,0.8012974,0.0006990182,0.0001609397,0.0001139483,0.0008234044,0.002955126,0.003382547],"genre_scores_gemma":[0.8508148,0.0006551487,0.1376966,0.0004512263,0.00008995576,0.0001885812,0.001571305,0.0001214763,0.008410945],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0168054,"threshold_uncertainty_score":0.0334152,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4233884501","doi":"10.1504/ijdmb.2020.109503","title":"Identification of protein hot regions by combining structure-based classification, energy-based clustering and sequence-based conservation in evolution","year":2020,"lang":"en","type":"article","venue":"International Journal of Data Mining and Bioinformatics","topic":"Protein Structure and Dynamics","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":"Simon Fraser University","funders":"","keywords":"Cluster analysis; Sequence (biology); Computer science; Identification (biology); Energy conservation; Data mining; Energy (signal processing); Pattern recognition (psychology); Artificial intelligence; Biological system; Algorithm; Mathematics; Statistics; Biology","authors":[{"name":"Jing Hu","is_ca":true},{"name":"Haomin Gan","is_ca":false},{"name":"Nansheng Chen","is_ca":true},{"name":"Xiaolong Zhang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0338256636374612,"gpt":0.276483070015979,"spread":0.2426574063785178,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005477826,0.0004541102,0.0006869494,0.002614077,0.0005804534,0.0005574376,0.0006472425,0.0005387908,0.0005255626],"category_scores_gemma":[0.00100513,0.0002378334,0.0006602096,0.001662751,0.0003106964,0.0008196099,0.0004839362,0.0003747474,0.0002612259],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003567995,"about_ca_system_score_gemma":0.0004441472,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003160438,"about_ca_topic_score_gemma":0.004128963,"domain_scores_codex":[0.9995149,0.00006771438,0.0000432937,0.0001733525,0.000143362,0.00005740661],"domain_scores_gemma":[0.9994596,0.000134015,0.0001322444,0.00005030949,0.0001734801,0.0000504341],"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":[0.000742592,0.0005444423,0.1570938,0.000520024,0.0005593566,0.0006260275,0.0005832225,0.07115608,0.1732711,0.004213929,0.003501982,0.5871875],"study_design_scores_gemma":[0.00003632972,0.0001793636,0.07088397,0.00004226799,0.0001750209,0.0007813202,0.0002251376,0.884883,0.03587299,0.004417936,0.002430466,0.00007220475],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6492838,0.001416913,0.3453829,0.0001941526,0.00004179118,0.0001584295,0.000377749,0.001134735,0.002009477],"genre_scores_gemma":[0.8523328,0.0003076407,0.145195,0.0000622935,0.00002410924,0.00006695327,0.0008520629,0.00007917672,0.001080009],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.003160438,"threshold_uncertainty_score":0.006284118,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}