{"id":"W4288917651","doi":"10.1007/978-3-031-01865-7_8","title":"Data Profiling Tools","year":2019,"lang":"en","type":"book-chapter","venue":"Synthesis lectures on data management","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Profiling (computer programming); Computer science; Visualization; Data science; Data mining; 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":[0.002694279,0.00186947,0.0009074848,0.005414724,0.001082666,0.006754798,0.002290463,0.001041451,0.08913841],"category_scores_gemma":[0.008739182,0.001352725,0.001229675,0.005773783,0.0006760074,0.008732699,0.004462269,0.002941071,0.08436187],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009761173,"about_ca_system_score_gemma":0.001477637,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001219632,"about_ca_topic_score_gemma":0.001588011,"domain_scores_codex":[0.9981504,0.0002586443,0.0001934212,0.0003759853,0.0009087806,0.0001126073],"domain_scores_gemma":[0.99657,0.00120849,0.0001366513,0.00116132,0.0007198441,0.000203749],"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":[0.00006058798,0.00005411053,0.0004516995,0.0004012494,0.00002607815,0.0001104076,0.0003111708,0.0009394155,0.003047558,0.1006321,0.3717552,0.5222105],"study_design_scores_gemma":[0.000008365812,0.000009193096,0.000252378,0.0001932956,0.00001305168,0.0002168384,0.00007844775,0.003496191,0.004085098,0.05196986,0.9396511,0.0000263342],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001251914,0.00217376,0.7588921,0.002243657,0.00122149,0.0003822406,0.01311352,0.06314607,0.1575753],"genre_scores_gemma":[0.01882453,0.004033499,0.5919562,0.002596924,0.0006615712,0.000679554,0.03641801,0.01903938,0.3257903],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.08913841,"threshold_uncertainty_score":0.2981977,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4881953596693671,"score_gpt":0.4276409553592674,"score_spread":0.06055440431009973,"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."}}