{"id":"W2899974865","doi":"10.1109/trpms.2018.2880617","title":"An Empirical Approach for Avoiding False Discoveries When Applying High-Dimensional Radiomics to Small Datasets","year":2018,"lang":"en","type":"article","venue":"IEEE Transactions on Radiation and Plasma Medical Sciences","topic":"Radiomics and Machine Learning in Medical Imaging","field":"Medicine","cited_by":25,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University Health Centre","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Terry Fox Research Institute","keywords":"Overfitting; Feature selection; Artificial intelligence; Radiomics; Logistic regression; Computer science; Support vector machine; Feature (linguistics); Machine learning; Receiver operating characteristic; Pattern recognition (psychology)","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001511868,0.0001673295,0.000265187,0.0002298291,0.0009513515,0.0001195343,0.000188962,0.0001279121,0.0001529492],"category_scores_gemma":[0.0002953491,0.0001271557,0.00005585171,0.0002933104,0.0006222657,0.0002075725,0.000003528602,0.0003752066,0.000008610921],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004912067,"about_ca_system_score_gemma":0.0002250584,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001203796,"about_ca_topic_score_gemma":0.00004223847,"domain_scores_codex":[0.998069,0.0001100806,0.0002973907,0.0005280302,0.0006540946,0.0003413395],"domain_scores_gemma":[0.9985178,0.0004573966,0.00005946786,0.0001642799,0.00003533097,0.0007656963],"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":[0.0008232918,0.0009110328,0.002437416,0.0001329693,0.0001241686,0.00001890485,0.00268662,0.0132938,0.002975248,0.001302977,0.007586543,0.967707],"study_design_scores_gemma":[0.001738989,0.001392494,0.0008724597,0.00007409602,0.00007960904,0.0001351095,0.0002453036,0.9816189,0.002868357,0.0002697118,0.01044746,0.000257501],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4007436,0.0000171505,0.592558,0.005757062,0.0003853198,0.0003647399,0.00006664346,0.00005017512,0.00005738072],"genre_scores_gemma":[0.9402468,0.00003263781,0.05465819,0.004379816,0.0004114274,0.00008511251,0.0001106386,0.00001477515,0.00006055828],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9683251,"threshold_uncertainty_score":0.7317123,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03991037103679596,"score_gpt":0.3378253655991552,"score_spread":0.2979149945623593,"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."}}