pub fn precision( predictions: &[usize], targets: &[usize], n_classes: usize, avg: Average, ) -> f64
Precision for multi-class classification.
Precision = TP / (TP + FP) — how many selected items are relevant.