Reproduction of IVFS algorithm for high-dimensional topology preservation feature selection

TL;DR

This paper reproduces the IVFS feature selection algorithm for high-dimensional data, confirming its effectiveness in preserving data topology and outperforming other methods, while noting some stability issues.

Contribution

It provides a systematic reproduction and validation of the IVFS algorithm, highlighting its strengths and limitations in topology-preserving feature selection.

Findings

IVFS outperforms SPEC and MCFS on most datasets

IVFS effectively preserves data topological structure

Convergence and stability issues remain in IVFS

Abstract

Feature selection is a crucial technique for handling high-dimensional data. In unsupervised scenarios, many popular algorithms focus on preserving the original data structure. In this paper, we reproduce the IVFS algorithm introduced in AAAI 2020, which is inspired by the random subset method and preserves data similarity by maintaining topological structure. We systematically organize the mathematical foundations of IVFS and validate its effectiveness through numerical experiments similar to those in the original paper. The results demonstrate that IVFS outperforms SPEC and MCFS on most datasets, although issues with its convergence and stability persist.