Precision

An evaluation metric for binary classification models that measures the proportion of samples predicted as positive that are truly positive, given by TP/(TP+FP); it typically exhibits a trade-off relationship with recall.