{"id":"W4225856506","doi":"10.1093/mutage/geac010","title":"Optimizing machine-learning models for mutagenicity prediction through better feature selection","year":2022,"lang":"en","type":"article","venue":"Mutagenesis","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Eisai Canada","keywords":"Machine learning; Artificial intelligence; Feature selection; Computer science; Benchmark (surveying); Context (archaeology); Cross-validation; Artificial neural network; Predictive modelling; Molecular descriptor; In silico; Model selection; Quantitative structure–activity relationship; Process (computing); Selection (genetic algorithm); Feature (linguistics); Data mining; Chemistry","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"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":[0.000729398,0.0001941598,0.0001939651,0.0001468244,0.001090486,0.0001958895,0.0005653661,0.00005799329,0.00004161616],"category_scores_gemma":[0.00004953541,0.0002175424,0.0001853845,0.0006957545,0.0000166459,0.0009673176,0.0004794708,0.0003200408,0.000003071118],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002767828,"about_ca_system_score_gemma":0.00009302584,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000603146,"about_ca_topic_score_gemma":0.000004093932,"domain_scores_codex":[0.9978744,0.0004591356,0.0002325412,0.0006233676,0.0004810451,0.0003294727],"domain_scores_gemma":[0.9991362,0.0003099789,0.000143275,0.0002621457,0.00009543867,0.00005294067],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002697703,0.00004538046,0.0002349681,0.00001599548,0.00004168615,0.000003168431,0.001214378,0.9815137,0.001180291,0.004169567,0.0006397816,0.01091411],"study_design_scores_gemma":[0.0003585123,0.0001251297,0.0003426103,0.000003518281,0.00002579748,0.00004046977,0.00005845169,0.9754841,0.002145801,0.009857964,0.01134976,0.0002078701],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04351348,0.000229042,0.9533014,0.001183149,0.0006021424,0.0003948184,0.00005665938,0.0003939254,0.0003254024],"genre_scores_gemma":[0.4419958,0.000006600284,0.5562241,0.0006704662,0.0001702295,0.0003283206,0.00007297906,0.00003298607,0.0004985086],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3984824,"threshold_uncertainty_score":0.887112,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03960028670017286,"score_gpt":0.2773120388961331,"score_spread":0.2377117521959602,"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."}}