{"id":"W4210979382","doi":"10.2200/s00878ed1v01y201810dtm052","title":"Data Profiling","year":2018,"lang":"en","type":"article","venue":"Synthesis lectures on data management","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":16,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Profiling (computer programming); Metadata; Computer science; Data science; Research data; Information retrieval; 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":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science","insufficient_payload"],"consensus_categories":["open_science","insufficient_payload"],"category_scores_codex":[0.01050815,0.000299099,0.000361493,0.0005002816,0.0004235989,0.00108646,0.01713791,0.00006541733,0.002094951],"category_scores_gemma":[0.006264756,0.0002154927,0.00005026725,0.0008559651,0.0002513714,0.001384135,0.01392284,0.0001553815,0.006231315],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004251727,"about_ca_system_score_gemma":0.0000267424,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004202726,"about_ca_topic_score_gemma":0.0003009929,"domain_scores_codex":[0.9933612,0.0004981734,0.0007871318,0.002385667,0.002424951,0.0005428895],"domain_scores_gemma":[0.9801961,0.001398888,0.000282326,0.01788284,0.00009968376,0.000140125],"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":[0.00006593727,0.0001253822,0.00004002249,0.00002171301,0.0001414978,0.00002230447,0.00002563823,0.00001218626,0.00001434532,0.02571108,0.590318,0.3835019],"study_design_scores_gemma":[0.0001551322,0.00005704506,0.0009592061,0.00004344177,0.0001025521,0.000001377058,0.00044311,0.003452102,0.0008636045,0.01156667,0.982058,0.0002977346],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.003240196,0.0003911623,0.2506053,0.02562048,0.004606259,0.003373264,0.01824465,0.0008294906,0.6930892],"genre_scores_gemma":[0.8634092,0.0003818785,0.0922583,0.01888594,0.002394791,0.0001739053,0.006347637,0.0001352229,0.01601309],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8601691,"threshold_uncertainty_score":0.9999505,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.484101918142008,"score_gpt":0.4815667505532124,"score_spread":0.002535167588795573,"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."}}