This article is cited in 2 papers

On a robust approach to search for cluster centers

Z. M. Shibzukhov<sup>ab

a</sup> Institute of Mathematics and Computer Science, Moscow Pedagogical State University, Moscow, 119991 Russia
^b Moscow Institute of Physics and Technology, Dolgoprudnyi, Moscow oblast, 141701 Russia

Abstract: We propose a new approach to the construction of $k$-means clustering algorithms in which the Mahalanobis distance is used instead of the Euclidean distance. The approach is based on minimizing differentiable estimates of the mean insensitive to outliers. Illustrative examples convincingly show that the proposed algorithm is highly likely to be robust with respect to a large amount of outliers in the data.

Keywords: cluster center, robust mean, Mahalanobis distance, iterative reweighting, robust algorithm.

Presented by the member of Editorial Board: A. A. Lazarev

Received: 24.01.2021
Revised: 26.04.2021
Accepted: 30.06.2021

DOI: 10.31857/S0005231021100111

 English version: Automation and Remote Control, 2021, 82:10, 1742–1751

Bibliographic databases:

• 
•