{"id":"W3126232929","doi":"10.1186/s13040-021-00244-z","title":"The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation","year":2021,"lang":"en","type":"article","venue":"BioData Mining","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":893,"is_retracted":false,"has_abstract":true,"ca_institutions":"Krembil Foundation","funders":"Deutsche Forschungsgemeinschaft","keywords":"Confusion matrix; Correlation; False positive paradox; Computer science; Confusion; False positives and false negatives; Classifier (UML); Statistics; Artificial intelligence; Metric (unit); Correlation coefficient; Machine learning; Data mining; Mathematics; Psychology; Operations management","routes":{"ca_aff":true,"ca_fund":false,"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":[0.001762428,0.0001970354,0.0002002522,0.0001260251,0.0003517381,0.0005833707,0.000739908,0.00009551262,0.00003247773],"category_scores_gemma":[0.0006715515,0.0001489541,0.00002999572,0.0006928727,0.0001055955,0.001234671,0.000704595,0.000183002,0.00002834772],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001623063,"about_ca_system_score_gemma":0.0003414543,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007134146,"about_ca_topic_score_gemma":0.00011563,"domain_scores_codex":[0.9977421,0.0001492541,0.0005368413,0.0005956199,0.0006270448,0.0003491307],"domain_scores_gemma":[0.9976253,0.0003633969,0.0003491719,0.001276071,0.0003178012,0.00006828783],"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":[0.0002876632,0.0003627121,0.1041116,0.0002320732,0.00007303976,0.00005586188,0.01516407,0.0007974996,0.02464442,0.03576079,0.09862348,0.7198868],"study_design_scores_gemma":[0.001067561,0.00003157498,0.05069922,0.0002263046,0.00001611916,0.00003875652,0.001250027,0.8883365,0.007416944,0.0003809202,0.05020287,0.0003331762],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5127193,0.002180492,0.4586311,0.01517173,0.002717055,0.003033906,0.0003182659,0.0008298899,0.004398221],"genre_scores_gemma":[0.9420449,0.0007553545,0.0531632,0.001355942,0.00007308814,0.0002808835,0.00098885,0.00003450177,0.001303316],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.887539,"threshold_uncertainty_score":0.607417,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02927417731283126,"score_gpt":0.3307727570661502,"score_spread":0.3014985797533189,"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."}}