Home | Publications | KHS19

On CoMADs and Principal Component Analysis

MCML Authors

Link to Profile Thomas Seidl

Thomas Seidl

Prof. Dr.

Director

Database Systems, Data Mining and AI

Abstract

Principal Component Analysis (PCA) is a popular method for linear dimensionality reduction. It is often used to discover hidden correlations or to facilitate the interpretation and visualization of data. However, it is liable to suffer from outliers. Strong outliers can skew the principal components and as a consequence lead to a higher reconstruction loss. While there exist several sophisticated approaches to make the PCA more robust, we present an approach which is intriguingly simple: we replace the covariance matrix by a so-called coMAD matrix. The first experiments show that PCA based on the coMAD matrix is more robust towards outliers.

inproceedings KHS19

SISAP 2019

12th International Conference on Similarity Search and Applications. Newark, New York, USA, Oct 02-04, 2019.

Authors

D. Kazempour • M. Hünemörder • T. Seidl

Links

DOI

Research Area

 A3 | Computational Models

BibTeXKey: KHS19

#p-seidl
Back to Top