Struct TimeSeriesSplit 

pub struct TimeSeriesSplit {
    pub n_splits: usize,
    pub mode: TimeSeriesMode,
    pub max_train_size: Option<usize>,
    pub gap: usize,
}

Time series cross-validation splitter.

Provides train/test indices for time series data where the test set is always in the future relative to the training set.

Example

use scirs2_core::data_split::{TimeSeriesSplit, TimeSeriesMode};

let ts = TimeSeriesSplit::new(3, TimeSeriesMode::Expanding, None).expect("ts");
let splits: Vec<_> = ts.split(20);
assert_eq!(splits.len(), 3);
// Training sets grow: each has more data than the previous

Fields

n_splits: usize

Number of splits

mode: TimeSeriesMode

Splitting mode

max_train_size: Option<usize>

Maximum training set size (only for Sliding mode)

gap: usize

Gap between train and test sets

Implementations

impl TimeSeriesSplit

pub fn new( n_splits: usize, mode: TimeSeriesMode, max_train_size: Option<usize>, ) -> CoreResult<Self>

Create a new TimeSeriesSplit.

pub fn with_gap(self, gap: usize) -> Self

Set the gap between train and test sets.

pub fn split(&self, n_samples: usize) -> Vec<SplitIndices>

Generate time series splits.

Trait Implementations

impl Clone for TimeSeriesSplit

fn clone(&self) -> TimeSeriesSplit

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TimeSeriesSplit

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.