Investigation of the optimal number of clusters by the adaptive EM algorithm

IF 0.5 Q4 ECONOMICS

Croatian Operational Research Review Pub Date : 2019-07-04 DOI:10.17535/CRORR.2019.0001

V. Novoselac

引用次数: 0

Abstract

This paper considers the investigation of the optimal number of clusters for datasets that are modeled as the Gaussian mixture. For that purpose, the adaptive method that is based on the modified Expectation Maximization (EM) algorithm is developed. The modification is conducted within the hidden variable of the standard EM algorithm. Assuming that data are multivariate normally distributed, where each component of the Gaussian mixture corresponds to one cluster, the modification is provided by utilizing the fact that the Mahalanobis distance of samples follows a Chi-square distribution. Besides, the quantity measure is constructed in order to determine number of clusters. The proposed method is presented in several numerical examples.

查看原文本刊更多论文

基于自适应EM算法的最优聚类数研究

本文考虑了高斯混合模型数据集的最优聚类数的研究。为此，提出了基于改进的期望最大化(EM)算法的自适应方法。修正是在标准电磁算法的隐变量内进行的。假设数据是多元正态分布的，其中高斯混合物的每个组成部分对应于一个簇，通过利用样本的马氏距离遵循卡方分布的事实来提供修改。此外，为了确定聚类的数量，构造了数量度量。给出了几个数值算例。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Croatian Operational Research Review ECONOMICS-

CiteScore

1.40

自引率

0.00%

发文量

审稿时长

22 weeks

期刊介绍： Croatian Operational Research Review (CRORR) is the journal which publishes original scientific papers from the area of operational research. The purpose is to publish papers from various aspects of operational research (OR) with the aim of presenting scientific ideas that will contribute both to theoretical development and practical application of OR. The scope of the journal covers the following subject areas: linear and non-linear programming, integer programing, combinatorial and discrete optimization, multi-objective programming, stohastic models and optimization, scheduling, macroeconomics, economic theory, game theory, statistics and econometrics, marketing and data analysis, information and decision support systems, banking, finance, insurance, environment, energy, health, neural networks and fuzzy systems, control theory, simulation, practical OR and applications. The audience includes both researchers and practitioners from the area of operations research, applied mathematics, statistics, econometrics, intelligent methods, simulation, and other areas included in the above list of topics. The journal has an international board of editors, consisting of more than 30 editors – university professors from Croatia, Slovenia, USA, Italy, Germany, Austria and other coutries.