vix.ing · top · new · best · stats · spec

Fuzzy Clustering to Identify Clusters at Different Levels of Fuzziness:\n An Evolutionary Multi-Objective Optimization Approach

2018/08/09 by Avisek Gupta, Shounak Datta, Gupta, Avisek +3
Computer Science · #Advanced Clustering Algorithms Research #Face and Expression Recognition

paper · pdf · doi:10.48550/arxiv.1808.03327

Abstract

Fuzzy clustering methods identify naturally occurring clusters in a dataset,\nwhere the extent to which different clusters are overlapped can differ. Most\nmethods have a parameter to fix the level of fuzziness. However, the\nappropriate level of fuzziness depends on the application at hand. This paper\npresents Entropy c-Means (ECM), a method of fuzzy clustering that\nsimultaneously optimizes two contradictory objective functions, resulting in\nthe creation of fuzzy clusters with different levels of fuzziness. This allows\nECM to identify clusters with different degrees of overlap. ECM optimizes the\ntwo objective functions using two multi-objective optimization methods,\nNon-dominated Sorting Genetic Algorithm II (NSGA-II), and Multiobjective\nEvolutionary Algorithm based on Decomposition (MOEA/D). We also propose a\nmethod to select a suitable trade-off clustering from the Pareto front.\nExperiments on challenging synthetic datasets as well as real-world datasets\nshow that ECM leads to better cluster detection compared to the conventional\nfuzzy clustering methods as well as previously used multi-objective methods for\nfuzzy clustering.\n

Related