{"id":"W2901302597","doi":"10.23889/ijpds.v3i5.1045","title":"Administrative Data Format Standardization for Efficient Analytics","year":2018,"lang":"en","type":"article","venue":"International Journal for Population Data Science","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Statistics Canada","funders":"","keywords":"Computer science; Standardization; Data quality; Data management; Data science; Data pre-processing; Metadata; Data governance; Data processing; Data warehouse; Database; Data mining; World Wide Web; Engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.03838347,0.001375087,0.001186456,0.007259388,0.002538135,0.01762154,0.005659533,0.002098405,0.01281604],"category_scores_gemma":[0.08196423,0.001345626,0.002038842,0.01616395,0.002458173,0.01544584,0.007769423,0.0061306,0.01382666],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004294088,"about_ca_system_score_gemma":0.01722725,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008941055,"about_ca_topic_score_gemma":0.004036108,"domain_scores_codex":[0.9677203,0.0103546,0.008663129,0.003275405,0.008834979,0.001151529],"domain_scores_gemma":[0.8789747,0.01460832,0.007438101,0.06359694,0.0339172,0.001464702],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000141441,0.000192477,0.004098576,0.000869564,0.00007554391,0.0002552285,0.00146279,0.006089345,0.004317205,0.6063263,0.1678336,0.2083379],"study_design_scores_gemma":[0.00003427167,0.00005705681,0.001702529,0.000621699,0.00004177261,0.0002374897,0.0006905815,0.007187074,0.008791556,0.05959982,0.9209387,0.00009739265],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00461244,0.001084614,0.8909556,0.0069747,0.002066764,0.002177421,0.01876505,0.01121278,0.06215078],"genre_scores_gemma":[0.04080975,0.002438914,0.8876351,0.002722685,0.0009777631,0.003025189,0.04311963,0.003653073,0.01561779],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03838347,"threshold_uncertainty_score":0.2029936,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5881116757594165,"score_gpt":0.5992713194419055,"score_spread":0.01115964368248901,"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."}}