pub fn recall<'a, I, T>(outputs: I, targets: I) -> f64where
I: ExactSizeIterator<Item = &'a T>,
T: 'a + PartialEq + Zero + One,
Expand description
Returns the recall score for 2 class classification.
Recall is calculated with true-positive / (true-positive + false-negative), see Precision and Recall for details.
Arguments
outputs
- Iterator of output (predicted) labels which only contains 0 or 1.targets
- Iterator of expected (actual) labels which only contains 0 or 1.
Examples
use rusty_machine::analysis::score::recall;
let outputs = [1, 1, 1, 0, 0, 0];
let targets = [1, 1, 0, 0, 1, 1];
assert_eq!(recall(outputs.iter(), targets.iter()), 0.5);
Panics
- outputs and targets have different length
- outputs or targets contains a value which is not 0 or 1