Specificity/ True Negative Rate:

It is common to compare sensitivity and specificity when assessing the performance of models. The percentage of true negatives the model correctly identifies is known as specificity. This suggests that a further proportion of true negatives - once thought to be positive could be referred to as false positives. This proportion may also be referred to as a True Negative Rate (TNR). The sum of specificity (true negative rate) and false positive rate would always equal one. A model with high specificity will accurately identify most negative results, whereas one with low specificity might mistakenly label several negative results as positive.

Let's utilise the model for determining whether a person has a disease.  Specificity is a measurement of the proportion of healthy individuals who were accurately identified as those who are not suffering from the disease. In other words, specificity is the proportion of those who were accurately predicted to be in good health.

Specificity can be mathematically calculated as follows:

Specificity = (True Negative)/(True Negative + False Positive)

The details on True Negative and False Positive used in the equation above are as follows:

True Negative: The number of people who are healthy and predicted to be healthy. In other words, the true negative is the number of people who are predicted to be healthy and who are actually healthy.

False Positive: People who were also predicted to be unhealthy or suffering from the disease turned out to be healthy. In other words, the number of people predicted to be unhealthy but actually healthy is represented by the false positive.

The model should ideally have a high specificity or true negative rate. A higher true negative value and a lower false-positive rate would result from a higher specificity value. A lower specificity value would result in a higher false positive and a lower true negative value.