{"id":"W2216178707","doi":"","title":"Big data curation","year":2014,"lang":"en","type":"article","venue":"International Conference on Management of Data","topic":"Big Data Technologies and Applications","field":"Decision Sciences","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Big data; Computer science; Data science; Data curation; Variety (cybernetics); Data management; Analytics; Semantics (computer science); Data modeling; Data mining; Database; 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":[],"consensus_categories":[],"category_scores_codex":[0.05907504,0.002766886,0.003887845,0.01717771,0.009908003,0.02588001,0.007742694,0.00504033,0.0252148],"category_scores_gemma":[0.1587323,0.001944106,0.004574188,0.02307143,0.007317612,0.0238953,0.02831701,0.009575178,0.02633414],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004783689,"about_ca_system_score_gemma":0.02654733,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00677852,"about_ca_topic_score_gemma":0.005811201,"domain_scores_codex":[0.8979953,0.03534688,0.01019308,0.01456688,0.03901517,0.002882638],"domain_scores_gemma":[0.7200003,0.06007205,0.01094908,0.1410538,0.05959969,0.008325146],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002079277,0.0001040741,0.003379725,0.005382971,0.0005676469,0.000655498,0.004989492,0.001378775,0.00547657,0.1852149,0.5280681,0.2645744],"study_design_scores_gemma":[0.00002507667,0.00001939672,0.0008098411,0.001022988,0.00006009232,0.0003487018,0.0006783555,0.001116472,0.00190239,0.05247514,0.9414722,0.00006929835],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004712134,0.0205094,0.7286547,0.04867149,0.01289907,0.004829557,0.02893122,0.02650213,0.1242903],"genre_scores_gemma":[0.05881045,0.02238513,0.7257776,0.03923267,0.01036417,0.006006308,0.07732506,0.01288619,0.04721244],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05907504,"threshold_uncertainty_score":0.3124224,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.7668386003949124,"score_gpt":0.4910349129574692,"score_spread":0.2758036874374432,"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."}}