| True positive |
TP |
|
| False positives (Type I error) |
FP |
|
| True negatives |
TN |
|
| False negatives (Type II error) |
FN |
|
| Accuracy |
(TP + TN) / (TP + FP + TN + FN) |
Unsuitable for unbalanced training sets |
| Sensitivity (True positive rate, Recall) |
TP / (TP + FN) |
|
| Specificity (True negative rate) |
TN / (TN + FP) |
|
| Positive predictive value (Precision) |
TP / (TP + FP) |
|
| Negative predictive value |
TN / (TN + FN) |
|
| Matthew’s correlation coefficient |
(TP × TN - FP × FN) / sqrt( (TP + FP) × (TP + FN) × (TN + FP) × (TN + FN) ) |
Suitable for unbalanced testing sets |
| Balanced error rate |
(FN/(TP + FN) + FP/(FP + TN)) / 2 |
Suitable for unbalanced testing sets |
| F-measure |
2 × (precision × recall)/(precision + recall) |
Does not take TN into account |
| Area under the curve |
|
Calculated from the ROC curve |