{"id":"W2574540012","doi":"10.1101/094888","title":"DataMed: Finding useful data across multiple biomedical data repositories","year":2016,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"U.S. National Library of Medicine; National Institute of Allergy and Infectious Diseases; National Human Genome Research Institute; University of California, San Diego; National Institute of Environmental Health Sciences; University of California, Los Angeles; University of Manchester; Yale University; European Bioinformatics Institute; Arizona State University; McGill University; National Institute of Diabetes and Digestive and Kidney Diseases; Johns Hopkins University; Purdue University; School of Medicine, New York University; National Institutes of Health; Universitair Medisch Centrum Groningen; GlaxoSmithKline","keywords":"Metadata; Interoperability; Computer science; Big data; Data science; Data discovery; World Wide Web; Knowledge extraction; Service (business); Reusability; Information retrieval; Data mining; Software","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02602119,0.001300063,0.001563381,0.01683413,0.002150838,0.0110932,0.004132461,0.002792916,0.005260518],"category_scores_gemma":[0.05696666,0.001130674,0.001854656,0.01300798,0.00212825,0.0121525,0.01228914,0.00204655,0.001970354],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002173916,"about_ca_system_score_gemma":0.004602063,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003260916,"about_ca_topic_score_gemma":0.005775979,"domain_scores_codex":[0.9805925,0.006507816,0.00228492,0.002883514,0.007146081,0.0005850916],"domain_scores_gemma":[0.9560319,0.02014085,0.004278083,0.01136805,0.005145709,0.003035336],"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":[0.002609648,0.001232367,0.07431273,0.004218296,0.001567794,0.002888677,0.002939449,0.01550617,0.03245621,0.09812188,0.158039,0.6061078],"study_design_scores_gemma":[0.0009672489,0.0008406056,0.02693763,0.002128472,0.0009148493,0.002856224,0.005547395,0.3293083,0.1116119,0.2417669,0.2764761,0.0006443755],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1226387,0.007295388,0.7239127,0.02996092,0.000989291,0.002171759,0.03467115,0.06270102,0.01565908],"genre_scores_gemma":[0.1981734,0.001050887,0.7676373,0.002012776,0.0002558307,0.0004334566,0.02600235,0.001664482,0.002769583],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02602119,"threshold_uncertainty_score":0.1376148,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2340761161406501,"score_gpt":0.4000028326354582,"score_spread":0.1659267164948081,"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."}}