The Benefits of the Matthews Correlation Coefficient (MCC) Over the Diagnostic Odds Ratio (DOR) in Binary Classification Assessment

摘要: To assess the quality of a binary classification, researchers often take advantage four-entry contingency table called confusion matrix , containing true positives, negatives, false and negatives. recap four values in unique score, statisticians have developed several rates metrics. In past, scientific studies already showed why Matthews correlation coefficient (MCC) is more informative trustworthy than confusion-entropy error, accuracy, F1 bookmaker informedness, markedness, balanced accuracy. this study, we compare MCC with diagnostic odds ratio (DOR), statistical rate employed sometimes biomedical sciences. After examining properties DOR, describe relationships between them, by also taking an innovative geometrical plot tetrahedron presented here for first time. We then report some use cases where DOR produce discordant outcomes, explain reliable two. Our results can strong impact computer science statistics, because they clearly trustworthiness information provided higher one generated ratio.