{"id":"W160971523","doi":"10.1007/s10230-014-0322-4","title":"On the Issue of Incomplete and Missing Water-Quality Data in Mine Site Databases: Comparing Three Imputation Methods","year":2014,"lang":"en","type":"article","venue":"Mine Water and the Environment","topic":"Groundwater flow and contamination studies","field":"Environmental Science","cited_by":16,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of British Columbia, Okanagan Campus; University of British Columbia","funders":"","keywords":"Missing data; Imputation (statistics); Water quality; Data mining; Data quality; Statistics; Multivariate statistics; Database; Computer science; Mathematics; 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":[0.002313427,0.0001289751,0.0002156208,0.00001902754,0.0001629021,0.00002701355,0.000191701,0.00001528995,0.0001332772],"category_scores_gemma":[0.0000205895,0.00005447963,0.00001611823,0.0000215209,0.0003957903,0.0001243161,0.001075873,0.00007135344,0.00002896866],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002510224,"about_ca_system_score_gemma":4.03289e-7,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001560477,"about_ca_topic_score_gemma":0.0003276328,"domain_scores_codex":[0.9987144,0.0003650798,0.0002743345,0.0002907143,0.0001837939,0.0001716429],"domain_scores_gemma":[0.9991854,0.0002671011,0.00004755639,0.000470869,0.000001389563,0.000027656],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003992618,0.0003904939,0.4436464,0.0001389344,0.0001530038,0.000004050085,0.02538644,0.002764456,0.2390687,0.002216591,0.0002475834,0.2855841],"study_design_scores_gemma":[0.003912302,0.0001399078,0.8239977,0.00004746544,0.0001129858,0.000006962111,0.0006274474,0.07342985,0.05425711,0.006835759,0.03618185,0.0004506548],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9334274,0.00005173271,0.06321937,0.002852422,0.00002439886,0.0001923861,0.000008868627,0.000004133113,0.0002193079],"genre_scores_gemma":[0.9974924,0.00001694387,0.00189909,0.0002754494,0.00001458689,0.00001312814,0.00008613716,0.000006391119,0.0001958532],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3803513,"threshold_uncertainty_score":0.2358985,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05431376854571816,"score_gpt":0.3045485893598897,"score_spread":0.2502348208141715,"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."}}