Constrained Sparse Subspace Clustering with Side-Information

TL;DR

This paper introduces CSSC+, an advanced subspace clustering method that effectively incorporates side-information during both affinity learning and spectral clustering, improving accuracy in high-dimensional data segmentation.

Contribution

It presents a novel constrained clustering approach that fully exploits side-information and provides a theoretical link to clustering accuracy, validated on gene expression data.

Findings

Enhanced clustering accuracy with side-information

Effective use of side-information in affinity and spectral clustering

Validated on cancer gene expression datasets

Abstract

Subspace clustering refers to the problem of segmenting high dimensional data drawn from a union of subspaces into the respective subspaces. In some applications, partial side-information to indicate "must-link" or "cannot-link" in clustering is available. This leads to the task of subspace clustering with side-information. However, in prior work the supervision value of the side-information for subspace clustering has not been fully exploited. To this end, in this paper, we present an enhanced approach for constrained subspace clustering with side-information, termed Constrained Sparse Subspace Clustering plus (CSSC+), in which the side-information is used not only in the stage of learning an affinity matrix but also in the stage of spectral clustering. Moreover, we propose to estimate clustering accuracy based on the partial side-information and theoretically justify the connection to the ground-truth clustering accuracy in terms of the Rand index. We conduct experiments on three cancer gene expression datasets to validate the effectiveness of our proposals.