Preprocessing splitting - CyrilB1531/lodestar GitHub Wiki

Development build. This page describes main, not a released package. The latest published Lodestar.Preprocessing is 0.1.0 — read its documentation.

Home › Preprocessing

Splitting — `Lodestar.Preprocessing`

One entry point, Splitters: it cuts the rows into cross-validation folds or into a single train and test split, at sklearn.model_selection parity wherever the reference is deterministic.

Indices in, indices out. A splitter here never sees the data. It is told how many rows there are — or, to stratify, what class each row belongs to — and it hands back the row numbers that train and the row numbers that are held out. Nothing is copied, and nothing decides the caller's layout.

Why this exists when ML.NET splits data

ML.NET has TrainTestSplit(IDataView, double, …) and CrossValidationSplit(IDataView, int, …), and both return data views. They also never stratify: dotnet/machinelearning#4396 has asked for it since 2019 and is open. samplingKeyColumnName keeps rows that share a key together, which is the opposite operation — it prevents a group from straddling the split, where stratifying spreads a class across every fold.

SharpLearning.CrossValidation does stratify, and its StratifiedIndexSampler<T> always shuffles from a seed, so it cannot reproduce a scikit-learn fold. What is missing in .NET is a splitter that is framework-free and reproducible, which is what decision 0004 wrote this for.

The permutation is an argument, not a seed

Each splitter has a second overload taking order, a permutation of 0..n−1 that it reads the rows in. Passing the permutation scikit-learn drew reproduces KFold(shuffle=True) and ShuffleSplit — the train/test split holds out the permutation's head, as ShuffleSplit does. Passing your own gives a split this package can describe exactly, without claiming a generator no reference shares — the same choice KMeansOptions.InitialCentres makes by taking the centres rather than a seed.

StratifiedKFold(shuffle=True) is the exception: it shuffles each class's fold list rather than the rows, so no permutation reproduces it, and the stratified order gives the unshuffled folds over the rows read in that order instead.

Types

Type	What it is
`Splitters`	The three splitters.
`FoldSplit`	One fold: which rows train, and which are held out.
`TrainTestSplit`	A single train and test split of the rows.