{"id":"W6958412261","doi":"10.6084/m9.figshare.25708975","title":"Additional file 1 of Ascertaining the Francophone population in Ontario: validating the language variable in health data","year":2024,"lang":"en","type":"article","venue":"Figshare","topic":"Census and Population Estimation","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Moncton; University of Manitoba; University of Ottawa; Institut du Savoir Montfort","funders":"","keywords":"Population; French; Variable (mathematics); Population health; Health data; Data collection","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":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.000242851,0.00006196991,0.00009972768,0.00005074928,0.00005934519,0.00003235517,0.0001670378,0.00003459439,0.9342163],"category_scores_gemma":[0.003491204,0.00004147284,0.00001728725,0.0003182053,0.000002425528,0.0001444305,0.00007694164,0.0002010531,0.00009722096],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009425592,"about_ca_system_score_gemma":0.0001792981,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01750544,"about_ca_topic_score_gemma":0.1003303,"domain_scores_codex":[0.9992141,0.00007501132,0.0003156871,0.0001291843,0.000160203,0.0001058395],"domain_scores_gemma":[0.9964507,0.003101818,0.0001324317,0.0002841382,0.00001897149,0.00001189718],"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":[0.000001373387,0.00001117696,0.00008585266,0.0001715309,0.000003334298,0.000001204662,0.001708185,0.0004467784,5.101522e-7,0.000252603,0.9952732,0.002044215],"study_design_scores_gemma":[0.0001982051,0.00001980017,0.1480755,0.01680694,0.000004170401,0.00001037069,0.001007129,0.2339613,0.00000231643,0.007043129,0.5926936,0.0001774947],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.0004543203,0.00005851719,0.00000811741,0.0001385512,0.00003616628,0.0002095492,0.9979894,0.00002402827,0.001081326],"genre_scores_gemma":[0.05148814,1.300934e-7,0.002943423,0.00006438885,0.00006178504,0.0001408415,0.9449427,0.00001105613,0.0003475666],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.9341191,"threshold_uncertainty_score":0.9890371,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1182249458262012,"score_gpt":0.3496066958957155,"score_spread":0.2313817500695142,"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."}}