{"id":"W2160183787","doi":"10.5210/ojphi.v7i1.5711","title":"Towards a Framework for Data Quality Properties of Indicators used in Surveillance","year":2015,"lang":"en","type":"article","venue":"Online Journal of Public Health Informatics","topic":"Data-Driven Disease Surveillance","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Bill and Melinda Gates Foundation","keywords":"Computer science; Data quality; Metadata; Data science; Disease surveillance; Quality (philosophy); Data mining; Data integration; Scalability; Population; Database; World Wide Web; Disease; Engineering; Medicine","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01385685,0.0001654852,0.001068624,0.0005578726,0.00002908407,0.00003353568,0.000711938,0.0001223995,0.000007971086],"category_scores_gemma":[0.01818282,0.000122732,0.00009162036,0.0007471982,0.0001208845,0.0008164724,0.0002000344,0.0004968049,0.000001896618],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003114341,"about_ca_system_score_gemma":0.008592678,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007915187,"about_ca_topic_score_gemma":0.0001165361,"domain_scores_codex":[0.9947596,0.0002959792,0.003392263,0.0001024252,0.0009929509,0.0004567606],"domain_scores_gemma":[0.9941655,0.0001986798,0.002925566,0.0008770907,0.0009697399,0.0008633554],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001142025,0.002855227,0.7213959,0.009970884,0.0003744401,0.00001102107,0.01465665,0.00006933766,0.00001082597,0.001863639,0.01816569,0.2294844],"study_design_scores_gemma":[0.01985252,0.003915355,0.5041879,0.004241684,0.00006337673,0.0002514625,0.02343784,0.0226124,0.00006445914,0.00129819,0.4192917,0.0007831321],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9292705,0.001484776,0.04474244,0.02089133,0.0004520521,0.000859252,0.002176782,0.00003415377,0.00008870581],"genre_scores_gemma":[0.8218925,0.000368298,0.1747894,0.002002134,0.0002653769,0.000005155526,0.0006419645,0.00002421851,0.0000109007],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.401126,"threshold_uncertainty_score":0.9970277,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3886151866498914,"score_gpt":0.4650183431601119,"score_spread":0.07640315651022045,"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."}}