Development build. This page describes main, not a released package. The latest published Lodestar.Preprocessing is 0.2.0 — read its documentation.

Home › Preprocessing › Splitting

Cuts the rows into contiguous folds, as KFold(n_splits=foldCount) does without shuffling.

public static IReadOnlyList<FoldSplit> KFold(int sampleCount, int foldCount)
public static IReadOnlyList<FoldSplit> KFold(int sampleCount, int foldCount, ReadOnlySpan<int> order)
public static IReadOnlyList<FoldSplit> KFold(int sampleCount, int foldCount, long randomState)

Parameters — sampleCount is how many rows there are. foldCount is how many folds to cut, at least two and at most sampleCount. order is a permutation of 0..sampleCount−1 the rows are read in; an empty span reads them in order, which is what the two-argument overload passes. randomState is scikit-learn's random_state, in [0, 2³² − 1]: it selects KFold(shuffle=True, random_state=randomState), numpy's generator replayed.

Returns — one FoldSplit per fold, in order, each index list ascending.

Exceptions — ArgumentOutOfRangeException when sampleCount is below two, or foldCount is below two or above sampleCount. ArgumentException when order is neither empty nor a permutation of the rows — a repeated index, an index out of range, or a length that is not sampleCount. ArgumentOutOfRangeException when randomState is outside numpy's range.

Example — ten rows into three folds, and the same ten read in a permuted order.

using Lodestar.Preprocessing;

IReadOnlyList<FoldSplit> folds = Splitters.KFold(sampleCount: 10, foldCount: 3);

// 10 = 3 + 3 + 3 + 1, and the leftover row goes to the first fold, not the last.
string first = string.Join(",", folds[0].TestIndices);   // => 0,1,2,3
string second = string.Join(",", folds[1].TestIndices);  // => 4,5,6
string third = string.Join(",", folds[2].TestIndices);   // => 7,8,9

// Reading the rows in another order cuts the same block sizes out of that order.
int[] order = [9, 4, 1, 7, 0, 3, 6, 8, 2, 5];
string shuffled = string.Join(",", Splitters.KFold(10, 3, order)[0].TestIndices);  // => 1,4,7,9

// scikit-learn's KFold(3, shuffle=True, random_state=42), fold for fold.
string seeded = string.Join(",", Splitters.KFold(10, 3, randomState: 42)[0].TestIndices);  // => 0,1,5,8

Remarks — the first sampleCount % foldCount folds take one extra row. That is the reference's placement and it is worth stating, because the other plausible rule — giving the remainder to the last folds — agrees with it only when the division is exact.

Every row is held out by exactly one fold and fitted on by the other foldCount − 1, permuted or not; that is asserted rather than assumed.

Applies to — net10.0, netstandard2.0.

See also — Splitters.StratifiedKFold, FoldSplit.