Error rate measures how often an inspection, classifier, or process produces an incorrect result. For classification, the misclassification rate is the number of false predictions divided by the total number of records: (false positives + false negatives) / n × 100. It is the inverse of accuracy, so accuracy + error rate is approximately 100%. section 5.2.4
For quality control, defect rate divides observed defects by the number of inspected units or other inspection opportunities. A c-chart tracks defect counts when each unit may contain multiple defects; a u-chart normalizes defects by the number of units, making samples of different sizes comparable. section 14.1.8
Raisin classification
A logistic regression classified 397 Kecimen observations correctly and 384 Besni observations correctly, with 53 false positives and 66 false negatives. Therefore, accuracy was 86.78% and the misclassification rate was 13.22%. In Python, calculate it as misclassification_rate = ((fp + fn) / n) * 100. Interpret the result alongside sensitivity, specificity, and AUC rather than treating it as a complete model assessment. section 5.2.4 section 5.3
Table 5.1 Confusion matrix for raisin classification. A confusion matrix stores classification metrics from a regression with actual and predicted binary outcomes.
|
Predicted
|
|||
|---|---|---|---|
|
Kecimen (0)
|
Besni (1)
|
||
|
Actual
|
Kecimen (0)
|
tn
|
fp
|
|
Besni (1)
|
fn
|
tp
|
|
Use the confusion matrix to identify true positives, true negatives, false positives, and false negatives before calculating error rate. section 5.2.4
