Clustering Three-Way Data with Outliers

TL;DR

This paper introduces an extension of the OCLUST algorithm for clustering matrix-variate normal data, effectively detecting and trimming outliers in complex structured data like images and time series.

Contribution

It presents a novel iterative method for outlier detection in matrix-variate normal models, addressing a gap in handling outliers in matrix-structured data.

Findings

Effective outlier detection in matrix-variate data

Extension of OCLUST algorithm to matrix data

Improved clustering robustness with outliers

Abstract

Matrix-variate distributions are a recent addition to the model-based clustering field, thereby making it possible to analyze data in matrix form with complex structure such as images and time series. Due to its recent appearance, there is limited literature on matrix-variate data, with even less on dealing with outliers in these models. An approach for clustering matrix-variate normal data with outliers is discussed. The approach, which uses the distribution of subset log-likelihoods, extends the OCLUST algorithm to matrix-variate normal data and uses an iterative approach to detect and trim outliers.