A Simple Approach to Sparse Clustering

TL;DR

This paper introduces a straightforward hill-climbing approach to sparse clustering, focusing on selecting relevant features, and demonstrates its competitiveness with existing methods like COSA and Sparse K-means.

Contribution

A simple hill-climbing algorithm for sparse clustering is proposed, offering an effective alternative to more complex existing methods.

Findings

The new method is competitive with COSA and Sparse K-means.

It simplifies the process of feature selection in clustering.

The approach is effective in identifying relevant features.

Abstract

Consider the problem of sparse clustering, where it is assumed that only a subset of the features are useful for clustering purposes. In the framework of the COSA method of Friedman and Meulman, subsequently improved in the form of the Sparse K-means method of Witten and Tibshirani, a natural and simpler hill-climbing approach is introduced. The new method is shown to be competitive with these two methods and others.