The Correct Matrix is ------------------------------Predicted Class Actual Class----[ 0 ]----------------[ 1 ] [ 0 ]---[π0 x specificity ]----------------[π0 x (1 - specificity)] [ 1 ]---[π1 x (1 - sensitivity)]----------------[ π1 x sensitivity ] Source: SAS Certification Prep Guide Statistical Business Analysis Using 9.4, December 2018, Shreve, Holland . Page 390 - Chapter 11 Measure of Model Performance

Mr. Mister's answer is WRONG. I literally just took this from the SAS video- Predictive Modeling Using Logistic Regression (15.1), LESSON 4.2 Predicted 0, Actual 0= pi0-specificity Predicted 1, Actual 0=pi0*(1-specificity) Predicted 0, Actual 1=pi1*(1-SENSITIVITY) Predicted 1, Actual 1= pi1*SENSITIVITY *Use specificity on actual False row, Use sensitivity on actual True row SOURCE: https://vle.sas.com/mod/scorm/player.php?a=5275&currentorg=ORG-EPMLR51&scoid=11485&newattempt=on

So correct confusion matrix would look like? Predicted 0 | 1 ------------------------------------------------------------------------------------ 0 | π0 x (1 - specificity) | π0 x specificity Actual -----------------------------------|------------------------------------- 1 | π1 x sensitivity | π1 x (1 - sensitivity) Just want to make sure I interpret your comment correctly. Thanks!

This is wrong. Actual Class 1 - should use pi1 / Actual class 0 - should use pi0