{"id":"W2748020237","doi":"10.1007/978-3-319-56994-9_78","title":"Corroborating Quality of Data Through Density Information","year":2017,"lang":"en","type":"book-chapter","venue":"Lecture notes in networks and systems","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":3,"is_retracted":false,"has_abstract":false,"ca_institutions":"McMaster University","funders":"","keywords":"Data deduplication; Tuple; Data quality; Computer science; Master data; Data mining; Data integration; Process (computing); Quality (philosophy); Data integrity; Data pre-processing; Database; Data science; Engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.007996664,0.0002575316,0.0009173933,0.000119718,0.0001895837,0.0007081376,0.001506328,0.0004344388,0.00002611135],"category_scores_gemma":[0.002447406,0.0001912201,0.00006404887,0.00005290399,0.0001790852,0.001414701,0.001165878,0.0004146702,0.00001545069],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002785694,"about_ca_system_score_gemma":0.00004539956,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001363723,"about_ca_topic_score_gemma":0.002884559,"domain_scores_codex":[0.9963828,0.0002321423,0.001659715,0.0005228286,0.001002172,0.0002003287],"domain_scores_gemma":[0.9930332,0.001889923,0.002141108,0.002638172,0.0002514465,0.00004610642],"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":[0.000169033,0.00003102616,0.002446462,0.0009723044,0.000189016,0.00001702246,0.002234129,0.06088384,0.000002251857,0.3624769,0.02266799,0.54791],"study_design_scores_gemma":[0.0007797088,0.00009065656,0.00130884,0.00172006,0.00008665398,0.00001258173,0.0002205402,0.1663815,0.000001860615,0.1773915,0.6510909,0.0009152021],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00007339572,0.003155498,0.9078392,0.0002151716,0.001640694,0.0006503126,0.0004831333,0.00002181718,0.08592074],"genre_scores_gemma":[0.9860928,0.00109123,0.001520422,0.0006334694,0.0008814512,0.000008801914,0.001939219,0.00002852714,0.007804048],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9860194,"threshold_uncertainty_score":0.7797727,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3017625718409327,"score_gpt":0.4222852155345378,"score_spread":0.120522643693605,"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."}}