« Back to Glossary Index

Precision is a performance metric applied to a classification problem and each class has its own value. For each class, the precision measures the capability of the model to apply the label of that class to an element that truly belongs to the class, i.e. it indicates from the predicted members of the class, how many are actually members. It equals the ratio of the number of correctly predicted instances of that class to the total number of predicted instances of that class.