{"id":"W4386207778","doi":"10.1080/19386389.2023.2251857","title":"From Uncontrolled Keywords to FAST? Attempting Metadata Reconciliation for a Canadian Research Data Aggregator","year":2023,"lang":"en","type":"article","venue":"Journal of Library Metadata","topic":"Research Data Management Practices","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Western University","funders":"","keywords":"Metadata; Discoverability; Computer science; World Wide Web; Workflow; Subject (documents); Metadata repository; News aggregator; Service (business); Brainstorming; Information retrieval; Database","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.07003118,0.0007553168,0.001116968,0.008065402,0.01503986,0.01526168,0.00439567,0.002635593,0.002209829],"category_scores_gemma":[0.1238203,0.001100874,0.00120007,0.01529172,0.008072901,0.01406289,0.01073476,0.003504627,0.001283153],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.02712925,"about_ca_system_score_gemma":0.09034216,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.453198,"about_ca_topic_score_gemma":0.5170751,"domain_scores_codex":[0.9558238,0.01601817,0.004216031,0.004748921,0.01601688,0.003176269],"domain_scores_gemma":[0.8825119,0.02557969,0.006221693,0.03523451,0.04710305,0.003349141],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0005023736,0.0002014484,0.03067769,0.0006975,0.0001577617,0.002308055,0.1167467,0.006668972,0.02244188,0.1237802,0.04190417,0.6539132],"study_design_scores_gemma":[0.0001042178,0.0002630808,0.01284772,0.0007394855,0.0002584738,0.001853367,0.09656644,0.031556,0.06483373,0.06552766,0.724772,0.0006779073],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3192467,0.002160365,0.5424919,0.06507403,0.001123535,0.002097481,0.00258523,0.01223104,0.0529897],"genre_scores_gemma":[0.3414865,0.0008013165,0.6335851,0.003837769,0.0001305033,0.0004779357,0.00191084,0.002059851,0.0157102],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9847383,"threshold_uncertainty_score":0.9011202,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4394956808097427,"score_gpt":0.4561766238239289,"score_spread":0.01668094301418621,"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."}}