{"id":"W3197767598","doi":"10.2196/31750","title":"Approaches and Criteria for Provenance in Biomedical Data Sets and Workflows: Protocol for a Scoping Review","year":2021,"lang":"en","type":"review","venue":"JMIR Research Protocols","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Bundesministerium für Bildung und Forschung; Deutsche Forschungsgemeinschaft","keywords":"Computer science; Workflow; Metadata; Data science; Data quality; Traceability; Transparency (behavior); Information retrieval; World Wide Web; Software engineering; Database; Engineering","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.1928644,0.006035444,0.01497186,0.02991634,0.007020758,0.0106297,0.00619525,0.01119935,0.04955658],"category_scores_gemma":[0.2454287,0.005712943,0.0167308,0.02992367,0.008710305,0.01134251,0.01105035,0.009043463,0.01193493],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01685229,"about_ca_system_score_gemma":0.09757268,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006510142,"about_ca_topic_score_gemma":0.009991553,"domain_scores_codex":[0.8535491,0.0530906,0.0669767,0.006314531,0.01502856,0.005040566],"domain_scores_gemma":[0.7692853,0.08711088,0.04043771,0.02557478,0.0736851,0.003906288],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"not_applicable","study_design_scores_codex":[0.004577565,0.0004192947,0.001436166,0.809516,0.002002766,0.001621838,0.008981642,0.001761185,0.002947578,0.01565099,0.07666736,0.07441766],"study_design_scores_gemma":[0.005872118,0.001124756,0.003588706,0.5598239,0.003136005,0.0007548251,0.005735278,0.001152519,0.003068414,0.01837354,0.3968708,0.0004991297],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"protocol","genre_gemma":"protocol","genre_scores_codex":[0.0004112522,0.00223233,0.003422901,0.0008263889,0.0006098122,0.9869993,0.004083185,0.0001142409,0.001300478],"genre_scores_gemma":[0.0002962925,0.0009547292,0.005698494,0.0001586003,0.00002734949,0.9920639,0.0005283937,0.00001140854,0.0002606648],"genre_candidate":"protocol","genre_consensus":"protocol","teacher_disagreement_score":0.8071356,"threshold_uncertainty_score":0.9953419,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.9494890316243164,"score_gpt":0.7658946204988578,"score_spread":0.1835944111254586,"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."}}