Izvorni znanstveni članak
https://doi.org/10.17535/crorr.2018.0017
Cluster Detection in Noisy Environment by Using the Modified EM Algorithm
Vedran Novoselac
; Mechanical Engineering Faculty in Slavonski Brod, J. J. Strossmayer University of Osijek, Slavonski Brod, Croatia
Zlatko Pavić
; Mechanical Engineering Faculty in Slavonski Brod, J. J. Strossmayer University of Osijek, Slavonski Brod, Croatia
Sažetak
The paper studies the problem of a cluster detection in the noisy environment. The solution of this problem is based on the well known Expectation Maximization (EM) algorithm. By utilizing the Mahalanobis distance, and modifying the hidden variable, the rejection procedure is constructed so that it omits data from calculation of the current iteration step. Thus we construct the adaptive framework for solving the above problem. Several numerical examples are presented to illustrate the proposed algorithm.
Ključne riječi
clustering; EM; Mahalanobis distance; least squares; least absolute devoation; Davies-Bouldin index
Hrčak ID:
212388
URI
Datum izdavanja:
13.12.2018.
Posjeta: 1.408 *