{"id":"W2177078353","doi":"10.4103/2153-3539.170649","title":"Extraction and analysis of discrete synoptic pathology report data using R","year":2015,"lang":"en","type":"article","venue":"Journal of Pathology Informatics","topic":"AI in cancer detection","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Parsing; Computer science; Scripting language; Quality assurance; Information retrieval; Data mining; Data extraction; Simple (philosophy); Natural language processing; Pathology; Medicine; Programming language; MEDLINE","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":[0.01653142,0.002151887,0.001368994,0.009097524,0.0008571363,0.003672859,0.002066487,0.0005160779,0.01975522],"category_scores_gemma":[0.09330335,0.001247331,0.002072121,0.007434563,0.001100347,0.001977432,0.002907324,0.001587099,0.0167871],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001207559,"about_ca_system_score_gemma":0.005942852,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003218523,"about_ca_topic_score_gemma":0.00320812,"domain_scores_codex":[0.9811506,0.005355549,0.005122643,0.003999791,0.003907979,0.0004634894],"domain_scores_gemma":[0.9171856,0.04514724,0.01040013,0.01484993,0.01126859,0.001148495],"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":[0.002578337,0.0003816887,0.04458687,0.006247975,0.001085587,0.001760935,0.003152766,0.0117466,0.04582508,0.01912012,0.1762065,0.6873075],"study_design_scores_gemma":[0.0007744641,0.00093527,0.09439218,0.001507665,0.0008025368,0.002041323,0.001963344,0.1595398,0.1599793,0.04923808,0.5280997,0.0007263929],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0210651,0.0001926751,0.7452355,0.0006703931,0.0001682209,0.003412772,0.09850936,0.1255439,0.005202018],"genre_scores_gemma":[0.07026757,0.0001980376,0.8507174,0.0001971525,0.0001168023,0.00458202,0.05978752,0.01093342,0.003200067],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01975522,"threshold_uncertainty_score":0.08742756,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08520372962763761,"score_gpt":0.3599386390515053,"score_spread":0.2747349094238677,"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."}}