Refer to the confusion matrix: Predicted Outcome 0 1 Actual Outcome 0 58 44 Actual Outcome 1 23 25 Calculate the accuracy and error rate (0 - negative outcome, 1 - positive outcome)

Question

Options

• AAccuracy = 58/102. Error Rate = 23/48
• BAccuracy = 83/102. Error Rate = 67/102
• CAccuracy = 25/150. Error Rate = 44/150
• DAccuracy = 83/150. Error Rate = 67/150

• A
  79% (15)
• B
  5% (1)
• C
  11% (2)
• D
  5% (1)

Explanation

There is an error in the provided answer key - the mathematically correct answer is actually D, not A.

Here's why: From the confusion matrix, you have TN=58, FP=44, FN=23, TP=25, giving a total of 150. Accuracy = (TP + TN) / Total = (25 + 58) / 150 = 83/150, and Error Rate = (FP + FN) / Total = (44 + 23) / 150 = 67/150 - which is option D.

Option A's values (58/102 and 23/48) are nonsensical: 102 is just the count of actual negatives (58+44), making 58/102 the specificity (true negative rate), not accuracy; and 23/48 is the false negative rate within the positive class - neither is overall accuracy or error rate.

Options B and C are wrong because: B uses 102 as the denominator (only actual-negative cases, not the full dataset), and C uses 25 and 44 which are individual cells, not the diagonal sum and off-diagonal sum.

Memory tip: Accuracy lives on the main diagonal (correct predictions: TN + TP) divided by the grand total. Error rate lives on the off-diagonal (wrong predictions: FP + FN) divided by the same grand total - and the two always sum to 1.

No community discussion yet for this question.