{"id":"W3211231734","doi":"10.5281/zenodo.4084763","title":"Data Repository Selection: Criteria That Matter","year":2020,"lang":"en","type":"article","venue":"Zenodo (CERN European Organization for Nuclear Research)","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"Victoria Park","funders":"","keywords":"Selection (genetic algorithm); Computer science; Artificial intelligence","routes":{"ca_aff":true,"ca_fund":false,"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":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.1837816,0.001663409,0.004565141,0.01839339,0.008462291,0.02929997,0.006209814,0.006999224,0.01380319],"category_scores_gemma":[0.4130566,0.001256615,0.00284873,0.02149394,0.004339524,0.01756877,0.008976453,0.003748516,0.01257019],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00443498,"about_ca_system_score_gemma":0.01861032,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003635001,"about_ca_topic_score_gemma":0.005604222,"domain_scores_codex":[0.7421916,0.1090989,0.05657029,0.01478574,0.06650315,0.01085028],"domain_scores_gemma":[0.4856595,0.1976532,0.03197295,0.05408777,0.2065517,0.02407499],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.003638642,0.001513272,0.0827554,0.01171193,0.001005507,0.00158392,0.007283936,0.002805618,0.01597288,0.05379903,0.2265273,0.5914025],"study_design_scores_gemma":[0.002284548,0.001423219,0.05854807,0.01203089,0.001751663,0.005075616,0.01776276,0.03308017,0.04368515,0.1131044,0.7102427,0.001010787],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.198532,0.02703485,0.4469619,0.1155711,0.009332297,0.03790362,0.03256859,0.008334162,0.1237614],"genre_scores_gemma":[0.4218051,0.003393646,0.4980712,0.01003075,0.002596394,0.0127906,0.02161923,0.004840314,0.02485272],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9937902,"threshold_uncertainty_score":0.9719415,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09740853218669542,"score_gpt":0.2794724886981486,"score_spread":0.1820639565114532,"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."}}