Biological and medical endeavors are beginning to realize the benefits of artificial intelligence and machine learning. However,\r\nclassification, prediction, and diagnostic (CPD) errors can cause significant losses, even loss of life.Hence, end users are best served\r\nwhen they have performance information relevant to their needs, this paper�s focus. Relative class size (rCS) is commonly recognized\r\nas a confounding factor in CPD evaluation. Unfortunately, rCS-invariant measures are not easily mapped to end user conditions.\r\nWe determine a cause of rCS invariance, joint probability table (JPT) normalization. JPT normalization means that more end user\r\nefficacious measures can be used without sacrificing invariance. An important revelation is that without data normalization, the\r\nMatthews correlation coefficient (MCC) and information coefficient (IC) are not relative class size invariants; this is a potential\r\nsource of confusion, as we found not all reports usingMCC or IC normalize their data.We deriveMCC rCS-invariant expression.\r\nJPT normalization can be extended to allow JPT rCS to be set to any desired value (JPT tuning). This makes sensitivity analysis\r\nfeasible, a benefit to both applied researchers and practitioners (end users).We apply our findings to two published CPD studies to\r\nillustrate how end users benefit.
Loading....