Random Manifold Sampling and Joint Sparse Regularization for Multi-label Feature Selection

TL;DR

This paper introduces a novel multi-label feature selection method combining random manifold sampling and joint sparse regularization, effectively addressing multicollinearity and improving feature relevance detection.

Contribution

It proposes a new model integrating $\, ext{l}_{2,1}$ and $\, ext{l}_F$ regularization with a random walk-based manifold regularization, along with an algorithm with proven convergence.

Findings

Outperforms existing methods on real-world datasets.

Effectively handles multicollinearity in feature selection.

Provides a robust neighborhood graph for better feature relevance.

Abstract

Multi-label learning is usually used to mine the correlation between features and labels, and feature selection can retain as much information as possible through a small number of features. $\ell_{2,1}$ regularization method can get sparse coefficient matrix, but it can not solve multicollinearity problem effectively. The model proposed in this paper can obtain the most relevant few features by solving the joint constrained optimization problems of $\ell_{2,1}$ and $\ell_{F}$ regularization.In manifold regularization, we implement random walk strategy based on joint information matrix, and get a highly robust neighborhood graph.In addition, we given the algorithm for solving the model and proved its convergence.Comparative experiments on real-world data sets show that the proposed method outperforms other methods.