The solution is to use measures that don't involve true negatives. Two such measures are as follows:
- Precision: This is the proportion of positive examples correctly predicted as positive (TP) out of all examples predicted as positive (TP + FP):
- Recall: This is the proportion of positives examples correctly predicted as positive (TP) out of all positive examples (TP + FN):
It is common to combine the two and report the F-measure, which considers both precision and recall to calculate the score as a weighted average, ...
