Revisiting Dimensionality Reduction Techniques for Visual Cluster Analysis: An Empirical Study

TL;DR

This study empirically evaluates how different dimensionality reduction techniques affect visual cluster analysis, highlighting the strengths of non-linear methods like UMAP and t-SNE for cluster identification.

Contribution

It provides a comprehensive user study comparing 12 DR techniques on multiple visual analysis tasks, revealing their relative effectiveness and user preferences.

Findings

Non-linear and local techniques are preferred for cluster and membership identification.

Linear techniques outperform in density comparison tasks.

UMAP and t-SNE excel in cluster and membership identification.

Abstract

Dimensionality Reduction (DR) techniques can generate 2D projections and enable visual exploration of cluster structures of high-dimensional datasets. However, different DR techniques would yield various patterns, which significantly affect the performance of visual cluster analysis tasks. We present the results of a user study that investigates the influence of different DR techniques on visual cluster analysis. Our study focuses on the most concerned property types, namely the linearity and locality, and evaluates twelve representative DR techniques that cover the concerned properties. Four controlled experiments were conducted to evaluate how the DR techniques facilitate the tasks of 1) cluster identification, 2) membership identification, 3) distance comparison, and 4) density comparison, respectively. We also evaluated users' subjective preference of the DR techniques regarding the quality of projected clusters. The results show that: 1) Non-linear and Local techniques are preferred in cluster identification and membership identification; 2) Linear techniques perform better than non-linear techniques in density comparison; 3) UMAP (Uniform Manifold Approximation and Projection) and t-SNE (t-Distributed Stochastic Neighbor Embedding) perform the best in cluster identification and membership identification; 4) NMF (Nonnegative Matrix Factorization) has competitive performance in distance comparison; 5) t-SNLE (t-Distributed Stochastic Neighbor Linear Embedding) has competitive performance in density comparison.