DataSplits.jl

DataSplits.jl provides a common interface for splitting data for model training, selection, and performance evaluation.

It includes random and resampling methods, cross-validation, group-aware and time-aware strategies, distance-based sample selection, and extrapolative splits.

Installation

using PkgPkg.add("DataSplits")

Quick start

using DataSplits# Diversity-based split — training set that covers the full feature space.res = partition(X, KennardStoneSplit(); train = 0.8, test = 0.2)X_train, X_test = splitdata(res, X)# Cover features and target jointly (SPXY).res = partition(X, SPXYSplit(); target = y, train = 80, test = 20)# Train / validation / test in one call.res = partition(X, RandomSplit(), KennardStoneSplit();                train = 70, validation = 10, test = 20)X_tr, X_val, X_te = splitdata(res, X)# Group-aware k-fold: no patient, scaffold, or batch spans two folds.cvs = partition(X, GroupKFold(5); groups = patient_ids)for (X_tr, X_te) in splitview(cvs, X)    fit!(model, X_tr)    evaluate(model, X_te)end

If you are new to data splitting, start with Why Splitting Matters.

For a hands-on introduction to the API, continue to Getting Started.