{"id":"W4210368113","doi":"10.2196/preprints.24381","title":"Automating Stroke Data Extraction From Free-Text Radiology Reports Using Natural Language Processing: Instrument Validation Study (Preprint)","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Acute Ischemic Stroke Management","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Medicine; Neurovascular bundle; Occlusion; Stroke (engine); Artificial intelligence; Radiology; Predictive value; Natural language processing; Nuclear medicine; Computer science; Surgery; Internal medicine","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.03340734,0.0007709299,0.0005494994,0.002006694,0.000597252,0.001368812,0.001129675,0.0006996225,0.001511757],"category_scores_gemma":[0.09479097,0.0003415153,0.001292488,0.001382964,0.0007851054,0.0008662765,0.001524948,0.0004962296,0.001446611],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008482032,"about_ca_system_score_gemma":0.002618462,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003613004,"about_ca_topic_score_gemma":0.002606877,"domain_scores_codex":[0.9804852,0.01167902,0.003102994,0.002516505,0.001796055,0.0004202406],"domain_scores_gemma":[0.8944782,0.0695073,0.005904729,0.008820435,0.02046581,0.000823527],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.004839246,0.008909835,0.5619302,0.002432604,0.001234792,0.0006186843,0.007029401,0.008673226,0.02534267,0.0005308831,0.009797405,0.368661],"study_design_scores_gemma":[0.002032502,0.007934631,0.8278018,0.000697918,0.001191517,0.001489024,0.003266068,0.08133466,0.05804455,0.0009548013,0.01499981,0.0002526942],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9612435,0.0003044811,0.02839042,0.0001737058,0.00007005709,0.003970008,0.004328579,0.0006810885,0.0008382527],"genre_scores_gemma":[0.8635126,0.0002904425,0.1116077,0.0003245872,0.00008991113,0.005404001,0.0177537,0.0001680062,0.0008490866],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03340734,"threshold_uncertainty_score":0.176677,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06504941291694595,"score_gpt":0.353581820252167,"score_spread":0.2885324073352211,"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."}}