Multilabel Feature Selection: A Local Causal Structure Learning Approach

Multilabel feature selection plays an essential role in high-dimensional multilabel learning tasks. Existing multilabel feature selection approaches mainly either explore the feature-label and feature-feature correlations or the label-label and feature-feature correlations. A few of them are able to deal with all three types of correlations simultaneously. To address this problem, in this article, we formulate multilabel feature selection as a local causal structure learning problem and propose a novel algorithm, M2LC. By learning the local causal structure of each class label, M2LC considers three types of feature relationships simultaneously and is scalable to high-dimensional datasets as well. To tackle false discoveries caused by the label-label correlations, M2LC consists of two novel error-correction subroutines to correct those false discoveries. Through local causal structure learning, M2LC learns the causal mechanism behind data, and thus, it can select causally informative features and visualize common features shared by class labels and specific features owned by an individual class label using the learned causal structures. Extensive experiments have been conducted to evaluate M2LC in comparison with the state-of-the-art multilabel feature selection algorithms.

Automated summary

This article presents a novel multilabel feature selection algorithm, M2LC, which addresses the correlations between features and labels through local causal structure learning. The algorithm considers three types of feature relationships simultaneously and corrects false discoveries through two error-correction subroutines. Experimental results demonstrate the effectiveness of M2LC in multilabel feature selection tasks.