{"id":"W4206522381","doi":"10.2196/32903","title":"Strategies to Address the Lack of Labeled Data for Supervised Machine Learning Training With Electronic Health Records: Case Study for the Extraction of Symptoms From Clinical Notes","year":2021,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Cancer Institute","keywords":"Artificial intelligence; Machine learning; Computer science; Task (project management); Bottleneck; Natural language processing; Information extraction; Named-entity recognition; Supervised learning; Set (abstract data type); Process (computing); Health records; Precision and recall; Health care; Artificial neural network","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":[0.004480456,0.0001667065,0.0005043354,0.00004987503,0.0003111663,0.0001078654,0.001409927,0.0001025958,0.00001989955],"category_scores_gemma":[0.002289461,0.00009458648,0.00007125742,0.0004013021,0.0001010012,0.0004238979,0.0004122163,0.0008979939,7.751247e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003688103,"about_ca_system_score_gemma":0.001998262,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001143557,"about_ca_topic_score_gemma":0.004626161,"domain_scores_codex":[0.9967687,0.0005042232,0.001350139,0.0002570659,0.0007004066,0.0004195051],"domain_scores_gemma":[0.9888481,0.008830852,0.0006530555,0.001163134,0.0003262096,0.0001786164],"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":[0.0004380668,0.0007609793,0.01485863,0.001348864,0.0006120176,0.00004533794,0.1509335,0.004196453,0.000007558655,0.004541848,0.0007271643,0.8215296],"study_design_scores_gemma":[0.001765869,0.002726424,0.001320968,0.0001995772,0.00004277865,0.0001397491,0.05689012,0.9330897,0.000009630197,0.0001102001,0.003573891,0.0001310688],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4870214,0.0002499574,0.5036899,0.006609734,0.0002173707,0.002037495,0.00009425601,0.00006576884,0.00001407499],"genre_scores_gemma":[0.9678942,0.00006294016,0.03048611,0.001082694,0.0001482774,0.0001884141,0.0001123795,0.00001676383,0.000008238385],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9288933,"threshold_uncertainty_score":0.3901387,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.177950069762978,"score_gpt":0.4690022669547648,"score_spread":0.2910521971917868,"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."}}