{"id":"W4403950719","doi":"10.23889/ijpds.v9i5.2932","title":"Leveraging Full Count Census Data through Record Linkage","year":2024,"lang":"en","type":"article","venue":"International Journal for Population Data Science","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadiana.org","funders":"","keywords":"Census; Record linkage; Computer science; Linkage (software); Count data; Data science; Statistics; Medicine; Population; Biology; Environmental health; Mathematics; Genetics","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":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.01532276,0.0001411858,0.0001659289,0.0005475529,0.0005890452,0.006394272,0.01329266,0.00003735159,0.0002897852],"category_scores_gemma":[0.007141498,0.0001083923,0.0000591337,0.0009020839,0.0001843709,0.01542157,0.003815959,0.0002493411,0.000212533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002076484,"about_ca_system_score_gemma":0.0002597428,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003685025,"about_ca_topic_score_gemma":0.0001689752,"domain_scores_codex":[0.9934773,0.0000912814,0.001028006,0.001162048,0.003918319,0.0003230149],"domain_scores_gemma":[0.9957987,0.0008056124,0.0003590944,0.002123329,0.0007804857,0.0001328],"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":[0.00007716733,0.00007376689,0.001241433,0.00001458058,0.00008035225,0.00007914997,0.0004234292,0.0006158901,0.0002594542,0.08077075,0.4325919,0.4837721],"study_design_scores_gemma":[0.0001550542,0.00002027643,0.002046688,0.00007015387,0.00001436759,0.00009137753,0.0003067065,0.1741634,0.000009177234,0.05533216,0.7676612,0.0001294928],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0213649,0.0004370221,0.9031435,0.02401344,0.03888318,0.0004152012,0.009441078,0.0001339062,0.002167734],"genre_scores_gemma":[0.8923285,0.0004633882,0.0896165,0.00245971,0.003379564,0.000008620402,0.008766144,0.00003182752,0.002945721],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8709636,"threshold_uncertainty_score":0.9983492,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5370824940918127,"score_gpt":0.5509539962253384,"score_spread":0.01387150213352562,"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."}}