EvoSplit: An evolutionary approach to split a multi-label data set into disjoint subsets

TL;DR

EvoSplit introduces an evolutionary algorithm to improve the partitioning of multi-label datasets into disjoint subsets, outperforming traditional methods like iterative stratification in maintaining label distributions.

Contribution

The paper presents a novel evolutionary approach, including single-objective and multi-objective algorithms, for better multi-label dataset splitting, validated on diverse datasets.

Findings

EvoSplit outperforms iterative stratification in preserving label distributions.

The multi-objective approach effectively balances label and label pair similarities.

Validated on both small and large datasets, including computer vision applications.

Abstract

This paper presents a new evolutionary approach, EvoSplit, for the distribution of multi-label data sets into disjoint subsets for supervised machine learning. Currently, data set providers either divide a data set randomly or using iterative stratification, a method that aims to maintain the label (or label pair) distribution of the original data set into the different subsets. Following the same aim, this paper first introduces a single-objective evolutionary approach that tries to obtain a split that maximizes the similarity between those distributions independently. Second, a new multi-objective evolutionary algorithm is presented to maximize the similarity considering simultaneously both distributions (labels and label pairs). Both approaches are validated using well-known multi-label data sets as well as large image data sets currently used in computer vision and machine learning applications. EvoSplit improves the splitting of a data set in comparison to the iterative stratification following different measures: Label Distribution, Label Pair Distribution, Examples Distribution, folds and fold-label pairs with zero positive examples.