Accuracy paradox

The accuracy paradox is the paradoxical finding that accuracy is not a good metric for predictive models when classifying in predictive analytics.

This is because a simple model may have a high level of accuracy but too crude to be useful.

Prior probabilities for these classes need to be accounted for in error analysis.

[citation needed] For example, a city of 1 million people has ten terrorists.

A profiling system results in the following confusion matrix: Even though the accuracy is ⁠10 + 999000/1000000⁠ ≈ 99.9%, 990 out of the 1000 positive predictions are incorrect.