Function cross_val_predict 

pub fn cross_val_predict<M: PipelineModel + Clone>(
    model: &M,
    data: &Dataset,
    k: usize,
    seed: u64,
) -> Result<Vec<f64>>

Out-of-fold predictions for every sample.

Trains on k-1 folds, predicts the held-out fold, and reassembles predictions in the original sample order. The returned vector has length data.n_samples().

Example

let preds = cross_val_predict(&model, &data, 5, 42)?;
assert_eq!(preds.len(), data.n_samples());