{"id":"W4407037890","doi":"10.2196/68863","title":"Performance Improvement of a Natural Language Processing Tool for Extracting Patient Narratives Related to Medical States From Japanese Pharmaceutical Care Records by Increasing the Amount of Training Data: Natural Language Processing Analysis and Validation Study","year":2025,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Computer science; Natural language processing; Natural (archaeology); Natural language; Narrative; Artificial intelligence; Linguistics; History","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008636495,0.001957038,0.000920247,0.002020121,0.0007651734,0.001220899,0.001267506,0.001511346,0.001439462],"category_scores_gemma":[0.01710457,0.0004616741,0.001372095,0.00112168,0.0005940645,0.002190246,0.001531953,0.001241448,0.0009717548],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001236962,"about_ca_system_score_gemma":0.001926756,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01320641,"about_ca_topic_score_gemma":0.009736927,"domain_scores_codex":[0.9947028,0.001816555,0.0008744122,0.001618373,0.0006020793,0.0003857957],"domain_scores_gemma":[0.9870723,0.008695534,0.0005282774,0.000964037,0.002347927,0.000391874],"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":[0.003775438,0.003204695,0.09488162,0.002074299,0.0013624,0.001459601,0.00203617,0.06535606,0.06596655,0.0004984175,0.01591811,0.7434667],"study_design_scores_gemma":[0.0002941238,0.002491444,0.06519333,0.0001513843,0.0005493984,0.000691132,0.001103387,0.8538571,0.07097505,0.0004656748,0.004085407,0.0001425736],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9556911,0.001404609,0.0333836,0.0004341775,0.0002138766,0.0003944711,0.002561823,0.004352158,0.001564217],"genre_scores_gemma":[0.904349,0.0004242982,0.07968722,0.0001957436,0.00005982685,0.0003548453,0.01320635,0.00009678736,0.001625894],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01320641,"threshold_uncertainty_score":0.04567468,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01543985965683919,"score_gpt":0.3759784832091741,"score_spread":0.3605386235523349,"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."}}