Multilabel Feature Selection: A Local Causal Structure Learning Approach
- 16 September 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 34 (6) , 3044-3057
Abstract
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.Keywords
Funding Information
- National Key Research and Development Program of China (2020AAA0106100)
- National Natural Science Foundation of China (61876206)
- Australian Research Council Discovery Projects (DP170101306)
This publication has 27 references indexed in Scilit:
- Multi-label feature selection based on max-dependency and min-redundancyNeurocomputing, 2015
- Fast multi-label feature selection based on information-theoretic feature rankingPattern Recognition, 2015
- A Tutorial on Multilabel LearningACM Computing Surveys, 2015
- A Review on Multi-Label Learning AlgorithmsIEEE Transactions on Knowledge and Data Engineering, 2013
- Feature selection for multi-label classification using multivariate mutual informationPattern Recognition Letters, 2013
- BASSUM: A Bayesian semi-supervised method for classification feature selectionPattern Recognition, 2011
- Partial Correlation Estimation by Joint Sparse Regression ModelsJournal of the American Statistical Association, 2009
- ML-KNN: A lazy learning approach to multi-label learningPattern Recognition, 2007
- Using Bayesian Networks to Analyze Expression DataJournal of Computational Biology, 2000
- Multiple Comparisons among MeansJournal of the American Statistical Association, 1961