Akansha YadavData Science & Coding (Python) | MySQL

Specificity in evaluation | specificity in machine learning

16 Apr 2024 · 0 min

Watch on YouTube

0:26

Akansha Yadav explains specificity, the metric that measures how well a model identifies actual negative instances. The formula is TN / (TN + FP), derived from the confusion matrix's four components: true positives, false negatives, true negatives, and false positives.

The another important parameter in the model evaluation is the specificity. Specificity is basically the coverage of the actual negative samples. So to calculate it, we need a confusion matrix, and then in the confusion matrix we will get 4 values: TP, FN, TN, and FP— true positive, false negative, true negative, and false positive. So what you have to do is true negative divided by true negative plus false positive.