{"id":"W4293162521","doi":"10.1007/978-3-031-01865-7","title":"Data Profiling","year":2019,"lang":"en","type":"book","venue":"Synthesis lectures on data management","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":31,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Profiling (computer programming); Metadata; Computer science; Data science; Research data; World Wide Web; Data curation; Operating system","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":[0.001977157,0.001393956,0.0008747936,0.00321533,0.001292031,0.005779337,0.001432788,0.0008742923,0.107385],"category_scores_gemma":[0.007005809,0.0009341888,0.0009055509,0.004117483,0.0006849014,0.007572649,0.003806139,0.002222118,0.1039382],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001147458,"about_ca_system_score_gemma":0.001544247,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001186638,"about_ca_topic_score_gemma":0.00149821,"domain_scores_codex":[0.9983326,0.0002298201,0.0001212636,0.0004104033,0.0008071041,0.00009879679],"domain_scores_gemma":[0.9973602,0.0005840572,0.00009150735,0.001109775,0.0006927178,0.0001618012],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00006758508,0.00002960635,0.0002584284,0.0002018853,0.00001243392,0.00005772099,0.0001284623,0.0008207094,0.002513938,0.07984988,0.380661,0.5353983],"study_design_scores_gemma":[0.0000043483,0.00001021645,0.0002097023,0.0001046344,0.000007887666,0.0001913082,0.00005374744,0.00239657,0.002767274,0.03897649,0.9552615,0.0000163267],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001671839,0.00421498,0.5438364,0.005721062,0.004155294,0.0005056391,0.00816022,0.02031301,0.4114216],"genre_scores_gemma":[0.02137296,0.004674035,0.2028699,0.003094672,0.001443534,0.0003620603,0.01313495,0.006407616,0.7466403],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.107385,"threshold_uncertainty_score":0.3592386,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4345880521667448,"score_gpt":0.4454418407113966,"score_spread":0.0108537885446518,"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."}}