{"id":"W1525672751","doi":"10.1186/1471-2105-6-75","title":"Ranking the whole MEDLINE database according to a large training set using text indexing","year":2005,"lang":"en","type":"article","venue":"BMC Bioinformatics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":58,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ontario Genomics","funders":"U.S. National Library of Medicine; Universität Stuttgart; Ontario Genomics Institute; Stem Cell Network; Ontario Genomics; Ontario Innovation Trust","keywords":"MEDLINE; Computer science; Information retrieval; Relevance (law); Ranking (information retrieval); Set (abstract data type); Controlled vocabulary; Vocabulary; Subject (documents); National library; Natural language processing; Artificial intelligence; World Wide Web; Library science; Linguistics","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006558745,0.002260725,0.002272421,0.01200453,0.002102422,0.002783946,0.002712545,0.00225872,0.004679297],"category_scores_gemma":[0.03719905,0.0006405343,0.001467754,0.007446938,0.0007812581,0.003825879,0.001983411,0.001191104,0.003920681],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001667558,"about_ca_system_score_gemma":0.002485624,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008930406,"about_ca_topic_score_gemma":0.01041051,"domain_scores_codex":[0.9949145,0.001276398,0.0008302169,0.00129267,0.001476531,0.0002096802],"domain_scores_gemma":[0.9768553,0.01514507,0.0007023145,0.0016398,0.005106782,0.0005507214],"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":[0.00129842,0.001289177,0.02195041,0.003265396,0.0004545027,0.001697367,0.0008803419,0.01971875,0.05736503,0.001250427,0.04160956,0.8492205],"study_design_scores_gemma":[0.0007037664,0.003594906,0.07097629,0.001146818,0.001774948,0.004933348,0.002136409,0.6957645,0.1402787,0.007595102,0.07074513,0.0003500781],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6029789,0.01529621,0.2663768,0.003211348,0.0007644142,0.003410679,0.02763062,0.06417789,0.01615319],"genre_scores_gemma":[0.4762374,0.003307368,0.433378,0.0008994208,0.0004282616,0.001346816,0.07880161,0.001097656,0.004503501],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.01200453,"threshold_uncertainty_score":0.03468639,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07540264134123546,"score_gpt":0.330097927578299,"score_spread":0.2546952862370636,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}