{"meta":{"query_hash":"5d90b1ae8224","filters":{"venue":"International Journal of Data Mining and Bioinformatics"},"cohort_total":20,"direct_labels_cover":1,"predictions_cover":20,"exported":20,"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/5d90b1ae8224","api":"https://metacan.xera.ac/api/v1/cohort?venue=International+Journal+of+Data+Mining+and+Bioinformatics"},"results":[{"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.08225493591224538,"score_gpt":0.40097271618555896,"score_spread":0.3187177802733136,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1975341743","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":[0.07370114,0.0007602567,0.9231595,0.00018602042,0.00006072383,0.00012708266,0.0002333608,0.0008847952,0.0008871445],"genre_scores_gemma":[0.5008396,0.0005018849,0.49678573,0.00011701207,0.000104431805,0.00025568245,0.0008390027,0.00012696927,0.00042976585],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99514186,0.0014750131,0.0005415822,0.0005130348,0.002157782,0.0001707748],"domain_scores_gemma":[0.99239105,0.0045631803,0.0007866278,0.0011912126,0.0009433257,0.0001246355],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0044952333,0.00064124283,0.0014115981,0.004881209,0.0005268084,0.00133446,0.0009923091,0.0006803724,0.00051425543],"category_scores_gemma":[0.015617882,0.00024372524,0.0010774969,0.0050559575,0.0006337696,0.002842846,0.001419049,0.0008681977,0.00029757238],"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":[0.000491652,0.00039942714,0.015883207,0.00084118923,0.0005792248,0.0003401512,0.0004970448,0.04937797,0.0969817,0.019482728,0.0026511713,0.8124745],"study_design_scores_gemma":[0.0001033821,0.00067804847,0.01997127,0.000101560574,0.00023509706,0.0013506772,0.00033120564,0.8168432,0.077578254,0.07335858,0.009302979,0.00014572887],"about_ca_topic_score_codex":0.0006492126,"about_ca_topic_score_gemma":0.0008487502,"teacher_disagreement_score":0.004881209,"about_ca_system_score_codex":0.000658806,"about_ca_system_score_gemma":0.00066668517,"threshold_uncertainty_score":0.023773372},"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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)","score_opus":0.11632856175214673,"score_gpt":0.3320784984421196,"score_spread":0.2157499366899729,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2018695726","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":[0.027520474,0.00029614472,0.97069037,0.00010621023,0.000023071332,0.00008900208,0.000049654183,0.000612796,0.00061226235],"genre_scores_gemma":[0.11819188,0.00024393808,0.88082606,0.000038310933,0.000010516826,0.00008472759,0.00011018486,0.00014174756,0.00035258092],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.998447,0.00029458705,0.00016223897,0.00036526017,0.000578143,0.00015290892],"domain_scores_gemma":[0.9981458,0.0007373497,0.0002718657,0.00025130835,0.00053010887,0.00006355342],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0020643903,0.0005868831,0.0011408298,0.0028009296,0.0006852791,0.0018339431,0.001069108,0.0013791905,0.0010412277],"category_scores_gemma":[0.0045740334,0.0006709107,0.0012957371,0.0016471533,0.0009560228,0.0014029966,0.0009927124,0.0010845207,0.0005679363],"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":[0.000360749,0.00014577851,0.0025411171,0.0006478387,0.00012632729,0.0002273878,0.0006410102,0.0549766,0.5786582,0.0054179276,0.0011656132,0.35509145],"study_design_scores_gemma":[0.000032182743,0.00026128438,0.007581606,0.000098028846,0.00010632866,0.00058104465,0.00023369452,0.6403051,0.33699667,0.009231878,0.00443806,0.00013404131],"about_ca_topic_score_codex":0.0016263013,"about_ca_topic_score_gemma":0.0019050445,"teacher_disagreement_score":0.0028009296,"about_ca_system_score_codex":0.00085124996,"about_ca_system_score_gemma":0.001151699,"threshold_uncertainty_score":0.010917664},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.04640594359035279,"score_gpt":0.3157167867357922,"score_spread":0.2693108431454394,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2045369746","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":[0.004886871,0.00014220994,0.9936372,0.00007450242,0.000009049665,0.000040792467,0.00004335215,0.0010294643,0.0001365653],"genre_scores_gemma":[0.084625676,0.00016207856,0.91386133,0.00012093791,0.000020819922,0.00013362334,0.00040184593,0.0001569438,0.0005167915],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9959681,0.0015517828,0.00022899038,0.000891448,0.0011932214,0.0001664827],"domain_scores_gemma":[0.9958182,0.0024298835,0.00026371767,0.00076053914,0.0006524499,0.0000752017],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005214515,0.0009775767,0.0012650983,0.0026689414,0.0005627051,0.0011457938,0.0013431049,0.001123429,0.0008134216],"category_scores_gemma":[0.0089701135,0.00047768946,0.0022382385,0.0022609453,0.0008955534,0.0011098947,0.0016261853,0.0014822687,0.00061642466],"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":[0.00023044217,0.00028806098,0.0055956715,0.00042797887,0.00039225133,0.0005173189,0.0003040446,0.25689813,0.04674844,0.01458054,0.001526684,0.6724904],"study_design_scores_gemma":[0.0000138834375,0.00011837724,0.0010972621,0.000024961368,0.00006785721,0.00021455843,0.000053688887,0.95165265,0.02757451,0.014987152,0.0041565755,0.000038600614],"about_ca_topic_score_codex":0.0025365998,"about_ca_topic_score_gemma":0.0024861016,"teacher_disagreement_score":0.005214515,"about_ca_system_score_codex":0.0006274102,"about_ca_system_score_gemma":0.0013618915,"threshold_uncertainty_score":0.027577281},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Categorical variable; Cluster analysis; Data type; Computer science; Data mining; Artificial intelligence; Machine learning","score_opus":0.10190478292986778,"score_gpt":0.35695398446423293,"score_spread":0.25504920153436517,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2063036182","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":[0.014177139,0.00027820998,0.9830005,0.00025581164,0.000039591137,0.0001327103,0.0005236165,0.00093672145,0.0006557521],"genre_scores_gemma":[0.113558084,0.0001591445,0.8808381,0.00025414806,0.000044324523,0.00044139303,0.0035087303,0.0002048828,0.0009911795],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9861137,0.0048648342,0.0014354355,0.0023585206,0.0045317938,0.0006956628],"domain_scores_gemma":[0.9808457,0.007942857,0.0015093823,0.0039037147,0.0053180316,0.00048030872],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008788549,0.0008889947,0.0016824435,0.0066573676,0.0016571253,0.004179042,0.0028796513,0.0019336712,0.0015851635],"category_scores_gemma":[0.038714204,0.00071442284,0.0024220126,0.007358713,0.0014938427,0.002877817,0.0042454153,0.0023634129,0.0016850083],"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":[0.0013851052,0.0004713289,0.04300907,0.0013763146,0.0010638266,0.0004983288,0.0025693749,0.1135197,0.041616682,0.07877936,0.013475291,0.7022356],"study_design_scores_gemma":[0.00008546578,0.0001684616,0.01270235,0.00016300753,0.00015592099,0.00060130475,0.0005239307,0.7944673,0.018504215,0.15719897,0.0152473645,0.00018172461],"about_ca_topic_score_codex":0.003568818,"about_ca_topic_score_gemma":0.00529055,"teacher_disagreement_score":0.008788549,"about_ca_system_score_codex":0.0020696248,"about_ca_system_score_gemma":0.0026203352,"threshold_uncertainty_score":0.046478868},"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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","score_opus":0.041370472827446304,"score_gpt":0.2592375920972098,"score_spread":0.21786711926976352,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2067203347","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":[0.0069268933,0.00019337653,0.9913408,0.00014808097,0.000015328982,0.00001732052,0.000110093875,0.00022584603,0.0010222285],"genre_scores_gemma":[0.78666353,0.0012932703,0.20499694,0.00014973673,0.00012831735,0.0005001894,0.00072348217,0.000098351025,0.005446253],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9993691,0.00022809817,0.00002664784,0.00020093114,0.00012863922,0.000046607325],"domain_scores_gemma":[0.99874866,0.00090379646,0.00013473137,0.000067632405,0.0001119984,0.0000330811],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009978717,0.0009593382,0.0010920293,0.0010148991,0.00046189132,0.0013464296,0.0017379805,0.0011855774,0.002027553],"category_scores_gemma":[0.0032837912,0.0005302363,0.0013162607,0.001128989,0.0012723694,0.0017867866,0.0008668565,0.0014850816,0.00043796835],"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":[0.00002290737,0.000017320339,0.00055561966,0.000028283346,0.00003200934,0.000038713722,0.000047148824,0.96108615,0.00055770227,0.030187316,0.00018632435,0.007240385],"study_design_scores_gemma":[0.0000018901876,0.0000045485,0.000041412295,0.0000020842413,0.000005183905,0.0000042092393,0.0000023275752,0.9924967,0.00007440796,0.007197274,0.00016686528,0.0000030082304],"about_ca_topic_score_codex":0.012266531,"about_ca_topic_score_gemma":0.0070604985,"teacher_disagreement_score":0.012266531,"about_ca_system_score_codex":0.0011166843,"about_ca_system_score_gemma":0.0010280772,"threshold_uncertainty_score":0.02439022},"labels":[],"label_agreement":null},{"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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","score_opus":0.0812831393055651,"score_gpt":0.35290580302199787,"score_spread":0.27162266371643273,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2072197604","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":[0.055899914,0.0008726976,0.93959326,0.00018893904,0.00005826837,0.00016698604,0.00038023773,0.002415626,0.0004239145],"genre_scores_gemma":[0.55944765,0.00030173728,0.4367423,0.0001324742,0.000077171586,0.00034967804,0.0018911117,0.000107193446,0.0009506618],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99828404,0.0005970905,0.00020134449,0.00027903574,0.000505943,0.00013255063],"domain_scores_gemma":[0.9964741,0.00178607,0.00030194913,0.00022405187,0.0011573822,0.000056438665],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038699645,0.0011254632,0.001712775,0.0021870492,0.00038281467,0.0007127302,0.0011290644,0.0009489756,0.0011899541],"category_scores_gemma":[0.007043,0.00023358554,0.0011984748,0.0013789968,0.00026086628,0.0008265266,0.0005196797,0.000976716,0.0007829675],"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":[0.0005633061,0.00045853932,0.008162682,0.00029478266,0.00028402824,0.00024893478,0.00006358231,0.04033889,0.032222208,0.0008083745,0.004709399,0.9118454],"study_design_scores_gemma":[0.000074748605,0.00033479466,0.00907948,0.000027155078,0.00011422013,0.0003077427,0.000043302283,0.96486,0.021965498,0.0013310342,0.0018160181,0.000045994268],"about_ca_topic_score_codex":0.0013088599,"about_ca_topic_score_gemma":0.0011809225,"teacher_disagreement_score":0.0038699645,"about_ca_system_score_codex":0.0002973044,"about_ca_system_score_gemma":0.00056010636,"threshold_uncertainty_score":0.020466566},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.033992309906928496,"score_gpt":0.30036973770142444,"score_spread":0.26637742779449597,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2103153231","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":[0.09663832,0.013174506,0.87408113,0.0009583925,0.0005673135,0.0009272879,0.0011069031,0.008806895,0.003739275],"genre_scores_gemma":[0.44412878,0.0019306628,0.54774076,0.0003596257,0.0006506378,0.00035852508,0.0023155569,0.00030801483,0.00220736],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9935336,0.0028403695,0.0004550014,0.0007245449,0.0021573193,0.0002890764],"domain_scores_gemma":[0.98792064,0.005457501,0.00079251267,0.0020770428,0.00345053,0.0003017814],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0039019126,0.001654339,0.0024693639,0.007882127,0.0013572276,0.0023830032,0.0019381053,0.0012554259,0.0024838191],"category_scores_gemma":[0.019519996,0.00059657067,0.0011014474,0.004752973,0.0005498593,0.0030594144,0.0016489963,0.001448547,0.0022589422],"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":[0.0008856748,0.00062733045,0.007416855,0.0009610887,0.00039233288,0.00061840913,0.0004072895,0.04025604,0.048155814,0.0058800937,0.0147647085,0.8796344],"study_design_scores_gemma":[0.00034121805,0.0016436676,0.014180141,0.0002160488,0.0006794527,0.0022817985,0.00059568795,0.8318621,0.07852088,0.040850688,0.028350107,0.00047823577],"about_ca_topic_score_codex":0.0035952863,"about_ca_topic_score_gemma":0.009043274,"teacher_disagreement_score":0.007882127,"about_ca_system_score_codex":0.0008095561,"about_ca_system_score_gemma":0.0016057808,"threshold_uncertainty_score":0.020635545},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.15965036198780436,"score_gpt":0.336622064281259,"score_spread":0.17697170229345466,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2119632600","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":[0.017730134,0.00033922965,0.98071367,0.00019522314,0.000021275782,0.000039800132,0.000022836948,0.00015847012,0.000779461],"genre_scores_gemma":[0.44819325,0.0003613713,0.5496056,0.0001474534,0.00006474879,0.00013194187,0.00011799017,0.00004503684,0.0013326305],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9974299,0.0010805001,0.00015294483,0.00047713786,0.00071801245,0.00014142696],"domain_scores_gemma":[0.9972101,0.0015210371,0.00047621864,0.00026771193,0.00039810222,0.00012680913],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036725071,0.0011809438,0.0014213292,0.001292674,0.00053804996,0.0020267766,0.0014879297,0.0012153855,0.0009410146],"category_scores_gemma":[0.0063004135,0.00053105317,0.001427984,0.0010603322,0.00093042117,0.002002161,0.0019030917,0.0011367677,0.0002177673],"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":[0.0004165865,0.00022259675,0.008367579,0.00037141392,0.0007628634,0.0005879054,0.00047544332,0.67912924,0.01897564,0.051246446,0.0009327946,0.23851143],"study_design_scores_gemma":[0.000016503056,0.00010071482,0.0004702723,0.000010193277,0.00008672701,0.00006332286,0.00003930089,0.98605806,0.0023138819,0.009810519,0.0010138226,0.00001660596],"about_ca_topic_score_codex":0.0013657432,"about_ca_topic_score_gemma":0.0018741542,"teacher_disagreement_score":0.0036725071,"about_ca_system_score_codex":0.0008634284,"about_ca_system_score_gemma":0.0009445988,"threshold_uncertainty_score":0.019422352},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.09114874666879506,"score_gpt":0.3597068872948939,"score_spread":0.26855814062609884,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2130013930","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":[0.04809575,0.00019522574,0.95016146,0.0000936255,0.000020402616,0.00004000044,0.00012511402,0.00096368475,0.0003047062],"genre_scores_gemma":[0.55195606,0.00021441461,0.44490054,0.00006795198,0.000063773274,0.00015052763,0.0010517379,0.00017192631,0.0014230707],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99898213,0.00034503074,0.0000769089,0.0002054083,0.00031332526,0.00007711013],"domain_scores_gemma":[0.99705505,0.0016794634,0.00024027795,0.00045079022,0.00046193352,0.00011243766],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010111217,0.00044724764,0.00086272735,0.0015502867,0.0003705683,0.00065336365,0.0008058471,0.000581492,0.0010143415],"category_scores_gemma":[0.004744747,0.00019863977,0.00069529243,0.0016919859,0.0005035076,0.0011311254,0.0007596402,0.0007604223,0.00048200155],"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":[0.00038149927,0.00026513857,0.0064053866,0.00016995965,0.00010885075,0.0002108648,0.00015969475,0.14221837,0.032494787,0.027064923,0.004309974,0.78621054],"study_design_scores_gemma":[0.000012505416,0.000060389742,0.0011755098,0.000004968413,0.000013674301,0.00009544084,0.000020542364,0.97755164,0.0036869561,0.01629074,0.0010756615,0.000012019931],"about_ca_topic_score_codex":0.0017470147,"about_ca_topic_score_gemma":0.0019894266,"teacher_disagreement_score":0.0017470147,"about_ca_system_score_codex":0.00045614433,"about_ca_system_score_gemma":0.0005192338,"threshold_uncertainty_score":0.005347371},"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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)","score_opus":0.04764101485050556,"score_gpt":0.3031798806265017,"score_spread":0.25553886577599616,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2146392361","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":[0.0020723352,0.00006187877,0.9972934,0.000061004666,0.00000985096,0.000014669001,0.000045609017,0.00035303505,0.00008807454],"genre_scores_gemma":[0.31031272,0.0005070503,0.684301,0.00020269016,0.00011053672,0.00046892767,0.0009036774,0.00027896598,0.0029143647],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9989831,0.0003768871,0.00006922643,0.00026183503,0.00022260718,0.00008640629],"domain_scores_gemma":[0.99607027,0.0030256524,0.00028456232,0.0002121853,0.00033823212,0.00006926606],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025002086,0.00080371974,0.0010905468,0.0006770665,0.00056108955,0.0010073569,0.0016550093,0.0012023012,0.0019380031],"category_scores_gemma":[0.00743671,0.0008511433,0.0012463151,0.00083062105,0.00084237213,0.0011797409,0.0011605285,0.0022673977,0.001114522],"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":[0.00018355322,0.000049441704,0.0008716657,0.00007635963,0.0000919488,0.000106016676,0.00009417582,0.9089769,0.004177429,0.012176313,0.00091277587,0.07228346],"study_design_scores_gemma":[0.0000040853874,0.000010165033,0.00005198423,0.0000025527645,0.0000049227037,0.000013310162,0.000002556548,0.9962025,0.0005021326,0.0030035307,0.00019364317,0.000008645838],"about_ca_topic_score_codex":0.012071884,"about_ca_topic_score_gemma":0.010316535,"teacher_disagreement_score":0.012071884,"about_ca_system_score_codex":0.0009744067,"about_ca_system_score_gemma":0.0017347346,"threshold_uncertainty_score":0.024003208},"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,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_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","score_opus":0.06555281677021664,"score_gpt":0.31880183367583853,"score_spread":0.2532490169056219,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2150689251","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":[0.0018136913,0.00008508334,0.9974848,0.00003020198,0.000012345273,0.000043123113,0.00006433063,0.0003349643,0.00013140317],"genre_scores_gemma":[0.05372161,0.00022653956,0.9440271,0.000069740345,0.000035221183,0.00025827257,0.00084051344,0.00013903824,0.00068200054],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9975535,0.0007873822,0.00017548342,0.00070565316,0.00065366103,0.00012422852],"domain_scores_gemma":[0.99650985,0.0017242488,0.00023376575,0.0005471893,0.0008835211,0.00010140111],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0033216274,0.0012281811,0.0013473834,0.003333425,0.0014276481,0.001181924,0.0029173293,0.0019046828,0.001695862],"category_scores_gemma":[0.011279674,0.0010145404,0.0020935147,0.0028524231,0.0006419956,0.0019758889,0.0012981675,0.0026800984,0.0015252859],"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":[0.00027170737,0.0002601013,0.006979713,0.00033583876,0.0005362267,0.00022330417,0.00049945427,0.3092438,0.011829305,0.037820354,0.005165458,0.6268347],"study_design_scores_gemma":[0.00001517308,0.000032386808,0.0008909398,0.000030140216,0.00005179515,0.00012846079,0.000036359193,0.9737394,0.0029397493,0.019708699,0.0023697657,0.00005710664],"about_ca_topic_score_codex":0.0118517745,"about_ca_topic_score_gemma":0.012965375,"teacher_disagreement_score":0.0118517745,"about_ca_system_score_codex":0.0012426162,"about_ca_system_score_gemma":0.0024211279,"threshold_uncertainty_score":0.02356559},"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":"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.043079755140257824,"score_gpt":0.3345422988701792,"score_spread":0.2914625437299214,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2891379954","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":[0.025723128,0.0006668029,0.9694735,0.00019187428,0.00005020312,0.00017212411,0.00037088824,0.0017981679,0.0015533118],"genre_scores_gemma":[0.3005241,0.0006594955,0.6957562,0.00007381535,0.000058217145,0.00034884113,0.0010862193,0.00011559426,0.0013775596],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99922776,0.00018181543,0.00006654881,0.00016949745,0.00029281116,0.00006157882],"domain_scores_gemma":[0.99900514,0.00046255832,0.00013566682,0.00006509024,0.00029250368,0.000039144325],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009986616,0.0009407758,0.0010339967,0.0042644786,0.00085410755,0.0006752654,0.0007665407,0.000822859,0.0010724764],"category_scores_gemma":[0.0026392697,0.0004062943,0.0010929847,0.0021679734,0.0003319865,0.0012240951,0.0005936473,0.0007686265,0.000479935],"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":[0.00026344138,0.0002889772,0.014523614,0.00039011473,0.0005342776,0.00019634928,0.00021477541,0.4946794,0.0074392357,0.008249397,0.0057898657,0.46743056],"study_design_scores_gemma":[0.00000962624,0.000027211616,0.0013423644,0.000011535271,0.000026337353,0.000052374286,0.00002553612,0.99144137,0.0015001313,0.0044896384,0.001055513,0.000018442703],"about_ca_topic_score_codex":0.014764791,"about_ca_topic_score_gemma":0.015187607,"teacher_disagreement_score":0.014764791,"about_ca_system_score_codex":0.000945556,"about_ca_system_score_gemma":0.0015333425,"threshold_uncertainty_score":0.029357672},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.028468735948894958,"score_gpt":0.31662038653124835,"score_spread":0.28815165058235337,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2946233355","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":[0.0056054075,0.00021382527,0.9919335,0.000112470465,0.00002722154,0.000062986874,0.00016951191,0.0011113307,0.00076378894],"genre_scores_gemma":[0.13481411,0.00036543654,0.8588982,0.00020336409,0.0001353079,0.0002976612,0.0012110827,0.00014882123,0.003926006],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9985948,0.00022370242,0.00009180169,0.0005419888,0.000453618,0.00009402604],"domain_scores_gemma":[0.9980598,0.00088999193,0.00016865379,0.00034143648,0.0004747275,0.00006527755],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011554918,0.0008720576,0.0013907262,0.0027534817,0.0006475269,0.0008743356,0.002025953,0.0015642114,0.002436496],"category_scores_gemma":[0.0039108167,0.00053924497,0.001589679,0.0018953453,0.0004477271,0.0011885931,0.00055704615,0.001203247,0.0013522878],"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":[0.00023982493,0.00034515135,0.0072607202,0.00029522838,0.00032070556,0.00035028058,0.00019506481,0.17597872,0.02779475,0.0108314175,0.006459001,0.7699291],"study_design_scores_gemma":[0.000032527416,0.00004843036,0.00089182553,0.00002073676,0.000054202457,0.00017192536,0.00001956006,0.9868608,0.0051339176,0.0043638153,0.0023774905,0.00002483544],"about_ca_topic_score_codex":0.01072119,"about_ca_topic_score_gemma":0.01468289,"teacher_disagreement_score":0.01072119,"about_ca_system_score_codex":0.0007970265,"about_ca_system_score_gemma":0.0011087044,"threshold_uncertainty_score":0.021317542},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.034353480761702686,"score_gpt":0.2755679535106321,"score_spread":0.24121447274892943,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3063921741","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":[0.66815114,0.001527966,0.32660592,0.00019258051,0.000040330626,0.00014580988,0.00037441275,0.0010793997,0.0018825532],"genre_scores_gemma":[0.8654767,0.00032793335,0.13205731,0.000058926675,0.0000202693,0.00006761984,0.00080520747,0.00007812653,0.0011079859],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995666,0.000060809285,0.000038555503,0.00015862104,0.0001248835,0.000050482427],"domain_scores_gemma":[0.9994785,0.00013426946,0.00012745541,0.000048765432,0.00016424122,0.000046843623],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00050348026,0.00042336565,0.0006303606,0.0025618076,0.0005900518,0.0005508239,0.0006546336,0.00053884165,0.0005545312],"category_scores_gemma":[0.0009780444,0.00023804254,0.0006200738,0.0016947412,0.00033759268,0.00079997646,0.00045225827,0.00036642587,0.00025697728],"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":[0.00073934416,0.00047451715,0.16985326,0.00054142275,0.0004985918,0.0006455273,0.0006973936,0.07214291,0.18600285,0.004991844,0.003177809,0.56023455],"study_design_scores_gemma":[0.000036963367,0.00018615786,0.08734952,0.00004868107,0.00017521885,0.0008958804,0.00028312067,0.86041975,0.0423507,0.005475211,0.0027005272,0.00007833522],"about_ca_topic_score_codex":0.0030938424,"about_ca_topic_score_gemma":0.004077433,"teacher_disagreement_score":0.0030938424,"about_ca_system_score_codex":0.0003594742,"about_ca_system_score_gemma":0.0004238314,"threshold_uncertainty_score":0.006151676},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.11575858728256372,"score_gpt":0.29292615442257136,"score_spread":0.17716756714000764,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3107387809","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":[0.18921293,0.0013546862,0.8012974,0.00069901824,0.0001609397,0.000113948256,0.0008234044,0.0029551263,0.0033825468],"genre_scores_gemma":[0.85081476,0.0006551487,0.1376966,0.00045122625,0.00008995576,0.00018858117,0.0015713053,0.00012147625,0.008410945],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99988127,0.000018095745,0.00000674096,0.000040455598,0.000027239774,0.000026071717],"domain_scores_gemma":[0.9998429,0.000054624656,0.000018778106,0.000013089866,0.00005436662,0.000016194594],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00043317722,0.00067906483,0.00064726255,0.0005445485,0.0002674568,0.00045838542,0.0016256123,0.0008081337,0.0013284822],"category_scores_gemma":[0.0005883135,0.00029916022,0.0005981874,0.00044322287,0.00023807115,0.00073618436,0.0005937798,0.0008871986,0.00031712189],"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":[0.00020932252,0.00016487754,0.004002255,0.000048821166,0.00013140283,0.00014231834,0.000025218036,0.86721134,0.0068985587,0.0027340604,0.0025694163,0.11586248],"study_design_scores_gemma":[0.0000026332534,0.00000747546,0.00008504629,8.739561e-7,0.0000038109686,0.000006339931,7.0004614e-7,0.9992656,0.000257105,0.0002529496,0.00011640613,0.0000012000794],"about_ca_topic_score_codex":0.016805395,"about_ca_topic_score_gemma":0.015988277,"teacher_disagreement_score":0.016805395,"about_ca_system_score_codex":0.0007413374,"about_ca_system_score_gemma":0.0010791961,"threshold_uncertainty_score":0.0334152},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.09629500041666537,"score_gpt":0.32095265186112765,"score_spread":0.22465765144446226,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3137172717","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":[0.007040991,0.000112712754,0.99117273,0.000056128985,0.000029285538,0.000080363185,0.00011869368,0.0010943551,0.00029468388],"genre_scores_gemma":[0.08291238,0.00018182116,0.9132122,0.00008925936,0.000045472545,0.00027695452,0.0010872461,0.00023457885,0.0019601462],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99858296,0.0002255945,0.00007556137,0.00052772515,0.00048027627,0.00010778694],"domain_scores_gemma":[0.99865794,0.0003630771,0.00018248193,0.00020796117,0.0004907191,0.00009774646],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012634835,0.0010640798,0.0008396587,0.0027689831,0.0009672371,0.00076330756,0.001467242,0.001011899,0.0012567341],"category_scores_gemma":[0.00343569,0.000505753,0.001418069,0.0024229058,0.00066206016,0.0013444807,0.0014798082,0.0012195818,0.00087478897],"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":[0.00057756965,0.00024397172,0.0055640778,0.00053387054,0.0004020885,0.00052707066,0.0006646224,0.12584774,0.1470859,0.008711333,0.007223864,0.7026179],"study_design_scores_gemma":[0.000035055058,0.00008561352,0.0030890875,0.000017818726,0.000055353634,0.00034103866,0.00010450587,0.9468786,0.035542417,0.006234701,0.0075483806,0.00006737581],"about_ca_topic_score_codex":0.0054917247,"about_ca_topic_score_gemma":0.009944635,"teacher_disagreement_score":0.0054917247,"about_ca_system_score_codex":0.0009033934,"about_ca_system_score_gemma":0.0012820025,"threshold_uncertainty_score":0.010919511},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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)","score_opus":0.09629500041666537,"score_gpt":0.32095265186112765,"score_spread":0.22465765144446226,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3214325870","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":[0.007040991,0.000112712754,0.99117273,0.000056128985,0.000029285538,0.000080363185,0.00011869368,0.0010943551,0.00029468388],"genre_scores_gemma":[0.08291238,0.00018182116,0.9132122,0.00008925936,0.000045472545,0.00027695452,0.0010872461,0.00023457885,0.0019601462],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99858296,0.0002255945,0.00007556137,0.00052772515,0.00048027627,0.00010778694],"domain_scores_gemma":[0.99865794,0.0003630771,0.00018248193,0.00020796117,0.0004907191,0.00009774646],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012634835,0.0010640798,0.0008396587,0.0027689831,0.0009672371,0.00076330756,0.001467242,0.001011899,0.0012567341],"category_scores_gemma":[0.00343569,0.000505753,0.001418069,0.0024229058,0.00066206016,0.0013444807,0.0014798082,0.0012195818,0.00087478897],"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":[0.00057756965,0.00024397172,0.0055640778,0.00053387054,0.0004020885,0.00052707066,0.0006646224,0.12584774,0.1470859,0.008711333,0.007223864,0.7026179],"study_design_scores_gemma":[0.000035055058,0.00008561352,0.0030890875,0.000017818726,0.000055353634,0.00034103866,0.00010450587,0.9468786,0.035542417,0.006234701,0.0075483806,0.00006737581],"about_ca_topic_score_codex":0.0054917247,"about_ca_topic_score_gemma":0.009944635,"teacher_disagreement_score":0.0054917247,"about_ca_system_score_codex":0.0009033934,"about_ca_system_score_gemma":0.0012820025,"threshold_uncertainty_score":0.010919511},"labels":[],"label_agreement":null},{"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.11575858728256372,"score_gpt":0.29292615442257136,"score_spread":0.17716756714000764,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4230716249","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":[0.18921293,0.0013546862,0.8012974,0.00069901824,0.0001609397,0.000113948256,0.0008234044,0.0029551263,0.0033825468],"genre_scores_gemma":[0.85081476,0.0006551487,0.1376966,0.00045122625,0.00008995576,0.00018858117,0.0015713053,0.00012147625,0.008410945],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99988127,0.000018095745,0.00000674096,0.000040455598,0.000027239774,0.000026071717],"domain_scores_gemma":[0.9998429,0.000054624656,0.000018778106,0.000013089866,0.00005436662,0.000016194594],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00043317722,0.00067906483,0.00064726255,0.0005445485,0.0002674568,0.00045838542,0.0016256123,0.0008081337,0.0013284822],"category_scores_gemma":[0.0005883135,0.00029916022,0.0005981874,0.00044322287,0.00023807115,0.00073618436,0.0005937798,0.0008871986,0.00031712189],"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":[0.00020932252,0.00016487754,0.004002255,0.000048821166,0.00013140283,0.00014231834,0.000025218036,0.86721134,0.0068985587,0.0027340604,0.0025694163,0.11586248],"study_design_scores_gemma":[0.0000026332534,0.00000747546,0.00008504629,8.739561e-7,0.0000038109686,0.000006339931,7.0004614e-7,0.9992656,0.000257105,0.0002529496,0.00011640613,0.0000012000794],"about_ca_topic_score_codex":0.016805395,"about_ca_topic_score_gemma":0.015988277,"teacher_disagreement_score":0.016805395,"about_ca_system_score_codex":0.0007413374,"about_ca_system_score_gemma":0.0010791961,"threshold_uncertainty_score":0.0334152},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.0338256636374612,"score_gpt":0.276483070015979,"score_spread":0.2426574063785178,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4233884501","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6492838,0.0014169135,0.34538293,0.0001941526,0.000041791176,0.0001584295,0.00037774903,0.0011347346,0.0020094775],"genre_scores_gemma":[0.85233283,0.0003076407,0.14519496,0.0000622935,0.000024109244,0.00006695327,0.0008520629,0.000079176716,0.0010800089],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99951494,0.00006771438,0.000043293705,0.00017335246,0.000143362,0.000057406607],"domain_scores_gemma":[0.99945956,0.00013401502,0.00013224443,0.000050309485,0.00017348012,0.000050434097],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005477826,0.00045411024,0.0006869494,0.002614077,0.00058045343,0.0005574376,0.0006472425,0.0005387908,0.00052556256],"category_scores_gemma":[0.0010051295,0.00023783339,0.00066020957,0.0016627508,0.0003106964,0.0008196099,0.00048393622,0.00037474738,0.00026122594],"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":[0.000742592,0.00054444233,0.15709378,0.00052002404,0.00055935665,0.00062602747,0.0005832225,0.07115608,0.17327113,0.0042139287,0.0035019822,0.5871875],"study_design_scores_gemma":[0.000036329722,0.0001793636,0.07088397,0.00004226799,0.00017502095,0.0007813202,0.00022513764,0.88488305,0.03587299,0.004417936,0.0024304662,0.00007220475],"about_ca_topic_score_codex":0.0031604383,"about_ca_topic_score_gemma":0.004128963,"teacher_disagreement_score":0.0031604383,"about_ca_system_score_codex":0.00035679952,"about_ca_system_score_gemma":0.00044414724,"threshold_uncertainty_score":0.0062841177},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_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","score_opus":0.043079755140257824,"score_gpt":0.3345422988701792,"score_spread":0.2914625437299214,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4252436249","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":[0.025723128,0.0006668029,0.9694735,0.00019187428,0.00005020312,0.00017212411,0.00037088824,0.0017981679,0.0015533118],"genre_scores_gemma":[0.3005241,0.0006594955,0.6957562,0.00007381535,0.000058217145,0.00034884113,0.0010862193,0.00011559426,0.0013775596],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99922776,0.00018181543,0.00006654881,0.00016949745,0.00029281116,0.00006157882],"domain_scores_gemma":[0.99900514,0.00046255832,0.00013566682,0.00006509024,0.00029250368,0.000039144325],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009986616,0.0009407758,0.0010339967,0.0042644786,0.00085410755,0.0006752654,0.0007665407,0.000822859,0.0010724764],"category_scores_gemma":[0.0026392697,0.0004062943,0.0010929847,0.0021679734,0.0003319865,0.0012240951,0.0005936473,0.0007686265,0.000479935],"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":[0.00026344138,0.0002889772,0.014523614,0.00039011473,0.0005342776,0.00019634928,0.00021477541,0.4946794,0.0074392357,0.008249397,0.0057898657,0.46743056],"study_design_scores_gemma":[0.00000962624,0.000027211616,0.0013423644,0.000011535271,0.000026337353,0.000052374286,0.00002553612,0.99144137,0.0015001313,0.0044896384,0.001055513,0.000018442703],"about_ca_topic_score_codex":0.014764791,"about_ca_topic_score_gemma":0.015187607,"teacher_disagreement_score":0.014764791,"about_ca_system_score_codex":0.000945556,"about_ca_system_score_gemma":0.0015333425,"threshold_uncertainty_score":0.029357672},"labels":[],"label_agreement":null}]}