{"id":"W2924905970","doi":"10.3390/geosciences9030139","title":"Subsampling of Regional-Scale Database for improving Multivariate Analysis Interpretation of Groundwater Chemical Evolution and Ion Sources","year":2019,"lang":"en","type":"article","venue":"Geosciences","topic":"Groundwater and Isotope Geochemistry","field":"Earth and Planetary Sciences","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université du Québec en Abitibi-Témiscamingue; Université du Québec à Chicoutimi","funders":"","keywords":"Groundwater; Multivariate statistics; Principal component analysis; Fluoride; Representativeness heuristic; Data mining; Aquifer; Computer science; Environmental science; Statistics; Geology; Chemistry; Mathematics","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":[0.0004119364,0.00007937148,0.0001809046,0.0001151897,0.00005866436,0.00003264015,0.000153377,0.00004014064,0.00008106967],"category_scores_gemma":[0.00002656299,0.00006011176,0.00006807545,0.000287384,0.0001533039,0.0003450669,0.00002062518,0.00003818699,0.000002048706],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000003418941,"about_ca_system_score_gemma":0.00002350623,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007030723,"about_ca_topic_score_gemma":0.0004791423,"domain_scores_codex":[0.9991201,0.00001945734,0.0002292467,0.000273323,0.0001919794,0.0001659407],"domain_scores_gemma":[0.999498,0.0001176231,0.0001516132,0.0001241915,0.00006523684,0.00004332876],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00007357446,0.00001645311,0.8152509,0.0001349402,0.00002385165,6.676663e-8,0.0006938225,0.0008955558,0.1789157,0.00001814939,0.00000193677,0.003975013],"study_design_scores_gemma":[0.0003223738,0.000143684,0.548613,0.00003866384,0.0001380916,0.000003125454,0.001429428,0.3879699,0.06063532,0.0005110929,0.00002791985,0.0001673438],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9718205,0.0001877363,0.02763161,0.00002213526,0.00008498347,0.0001244284,0.00005582943,0.000007778689,0.00006500749],"genre_scores_gemma":[0.9960452,0.00000514023,0.003698349,0.000009644177,0.00002214024,0.000001370019,0.0001674361,0.000001180148,0.00004949175],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3870744,"threshold_uncertainty_score":0.9995815,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01036571594853353,"score_gpt":0.2174497358543444,"score_spread":0.2070840199058109,"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."}}