{"id":"W4320065330","doi":"10.2196/43014","title":"Extracting Medical Information From Free-Text and Unstructured Patient-Generated Health Data Using Natural Language Processing Methods: Feasibility Study With Real-world Data","year":2023,"lang":"en","type":"article","venue":"JMIR Formative Research","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":59,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Center for Advancing Translational Sciences; National Institutes of Health","keywords":"Computer science; Natural language processing; Pipeline (software); Artificial intelligence; Named-entity recognition; Unstructured data; Sentence; SNOMED CT; Parsing; Set (abstract data type); Unified Medical Language System; Information retrieval; Big data; Terminology; Data mining; Linguistics","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01049431,0.0002274147,0.0003919691,0.0006231286,0.001059783,0.0007504015,0.003415456,0.00007152381,0.0000131598],"category_scores_gemma":[0.002364947,0.0001706839,0.00001269993,0.003315365,0.000208582,0.006316517,0.006942576,0.001828917,0.00000856425],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002849166,"about_ca_system_score_gemma":0.001238752,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01709524,"about_ca_topic_score_gemma":0.007056592,"domain_scores_codex":[0.9915335,0.003816925,0.0007644925,0.0008273656,0.002257238,0.0008004843],"domain_scores_gemma":[0.9944499,0.001308703,0.000484218,0.002933035,0.0004726949,0.0003514301],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00009450963,0.00008092931,0.01830579,0.0002962418,0.00003307637,0.00002458274,0.07219907,0.00004051882,0.00002555238,0.0000451112,0.0006309283,0.9082237],"study_design_scores_gemma":[0.0006535982,0.0002982391,0.09009396,0.0001941268,0.00000278845,0.0000147791,0.01602838,0.8922396,0.00002267141,0.00008848978,0.0001915447,0.0001717868],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9599583,0.0003423416,0.03444108,0.002314488,0.0002019896,0.001913508,0.0003130618,0.0004164865,0.00009871428],"genre_scores_gemma":[0.9133088,0.00001575596,0.08513594,0.0001072828,0.00006979016,0.00002760582,0.001312278,0.00001713609,0.000005426064],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9080519,"threshold_uncertainty_score":0.98945,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2386185537144399,"score_gpt":0.5579463831016811,"score_spread":0.3193278293872412,"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."}}