{"id":"W4401823118","doi":"10.3233/shti240511","title":"Synthetic Generation of Patient Service Utilization Data: A Scalability Study","year":2024,"lang":"en","type":"article","venue":"Studies in health technology and informatics","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Victoria","funders":"","keywords":"Scalability; Computer science; Data science; Health services; Transformer; Data format; Synthetic data; Service (business); Data as a service; Data modeling; Data mining; Artificial intelligence; Database; Medicine; Engineering","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01237473,0.0005110608,0.000501373,0.0009035794,0.0003966971,0.0008339508,0.001296813,0.0007024542,0.00116273],"category_scores_gemma":[0.04448482,0.0002300685,0.0005484978,0.001241127,0.0008416393,0.001439559,0.001010771,0.001094467,0.0001855209],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001677057,"about_ca_system_score_gemma":0.001231683,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01608814,"about_ca_topic_score_gemma":0.009790253,"domain_scores_codex":[0.9960472,0.002591365,0.0001999018,0.0004050619,0.0006169125,0.0001394353],"domain_scores_gemma":[0.9520004,0.03778979,0.001159152,0.005807023,0.002804512,0.0004390898],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004909917,0.0003801424,0.03006814,0.0001641636,0.0001759999,0.0001651929,0.0002430891,0.9146305,0.001694051,0.005040944,0.005033464,0.04191328],"study_design_scores_gemma":[0.00004018783,0.000101515,0.003130439,0.00001185987,0.00001320908,0.0000633946,0.000101668,0.9918188,0.001330748,0.002686897,0.0006913278,0.000009974239],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9108109,0.0006193814,0.07699776,0.001929895,0.0001282896,0.0004171487,0.004106807,0.001262886,0.003726797],"genre_scores_gemma":[0.9692997,0.0001306311,0.0256301,0.0001747345,0.00002487009,0.0001167945,0.004186123,0.00005368543,0.0003833881],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01608814,"threshold_uncertainty_score":0.06544459,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1943336483599897,"score_gpt":0.4025065384496941,"score_spread":0.2081728900897044,"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."}}