Evolutionary static and dynamic clustering algorithms based on multi-verse optimizer. (June 2018)

• Record Type: Journal Article
• Authors: Shukri, Sarah; Faris, Hossam; Aljarah, Ibrahim; Mirjalili, Seyedali; Abraham, Ajith
• Is Part Of: Engineering applications of artificial intelligence. Volume 72(2018)
• Page Start: 54
• Page End: 66
• Publication Date: 2018-06
• Subjects: Optimization -- Metaheuristics -- Multi-verse optimizer -- Clustering -- Swarm Intelligence -- Data mining -- Machine learning -- Evolutionary Computation; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2018.03.013
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Clustering based on nature-inspired algorithms is considered as one of the fast growing areas that aims to benefit from such algorithms to formulate a clustering problem as an optimization problem. In this work, the search capabilities of a recent nature-inspired algorithm called Multi-verse Optimizer (MVO) is utilized to optimize clustering problems in two different approaches. The first one is a static clustering approach that works on a predefined number of clusters. The main objective of this approach is to maximize the distances between different clusters and to minimize the distances between the members in each cluster. In an attempt to overcome one of the major drawbacks of the traditional clustering algorithms, the second proposed approach is a dynamic clustering algorithm, in which the number of clusters is automatically detected without any prior information. The proposed approaches are tested using 12 real and artificial datasets and compared with several traditional and nature-inspired based clustering algorithms. The results show that static and dynamic MVO algorithms outperform the other clustering techniques on the majority of datasets.