Question

The following confusion matrix for a classification method with Class 0 and Class 1 was found by Data Mining: Confusion Matrix Predicted Class Actual Class 0 1 0 105 53 1 32 110 What are the percent error rates for Class 0 and Class 1? a. 31.65% and 22.54%, respectively b. 33.54% and 22.54%, respectively c. 10.71% and 41.25%, respectively d. 45.37% and 46.67%, respectively

Answer 1

b. 33.54% and 22.54%, respectively

Explanation:

Confusion Matrix

Predicted Class

Actual Class 0 1 Total

0 105 53 158

True Positive False Negative

TP FN

1 32 110 142

False Positive True Negative

FP TN

Total Error Rate = FN +FP/ P +N

Total Error Rate = 53+32/ 105+32+53+110

Total Error Rate = 85/300 = 28.33%

Error Rate for Class 0 = FP/ FP+TP= 32/142 = 22.53%

Error Rate for Class 1= FN/ FN+TP= 53/ 158= 33.54%