{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01584418,0.001545204,0.001511071,0.007663949,0.001366348,0.003296727,0.001054969,0.00191533,0.001817628],"category_scores_gemma":[0.1107636,0.0003350041,0.0008541324,0.004399135,0.002304319,0.002821813,0.001821487,0.001428677,0.0009953736],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001400318,"about_ca_system_score_gemma":0.001319449,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003817915,"about_ca_topic_score_gemma":0.005058927,"domain_scores_codex":[0.9814667,0.005495087,0.002551811,0.003180954,0.0066273,0.0006782635],"domain_scores_gemma":[0.8398484,0.1025892,0.02223467,0.008470453,0.02368741,0.003169879],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0024924,0.0003610173,0.4226277,0.001822262,0.00262621,0.0009373054,0.002193426,0.1344294,0.0186979,0.02001596,0.04351469,0.3502817],"study_design_scores_gemma":[0.00009906881,0.001587682,0.1799962,0.0005762462,0.000605116,0.001548274,0.001153974,0.7163308,0.04192943,0.04094587,0.01463181,0.0005954534],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6303577,0.007464869,0.3337457,0.00225208,0.0009788257,0.0005293683,0.003544639,0.004319874,0.01680693],"genre_scores_gemma":[0.9578433,0.0002957901,0.03816509,0.0002443196,0.0002563609,0.0001537767,0.001516277,0.0002316614,0.001293396],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.01584418,"threshold_uncertainty_score":0.08379304,"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."}}