EvoCluster: An Open-Source Nature-Inspired Optimization Clustering Framework

• Published in: SN computer science Vol. 2; no. 3; p. 185
• Main Authors: Qaddoura, Raneem, Faris, Hossam, Aljarah, Ibrahim, Castillo, Pedro A.
• Format: Journal Article
• Language: English
• Published: Singapore Springer Singapore 01.05.2021; Springer Nature B.V
• Subjects: C plus plus; Clustering; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data Structures and Information Theory; Datasets; Evolution; Genetic algorithms; Heuristic methods; Information Systems and Communication Service; Libraries; Open source software; Optimization algorithms; Original Research; Pattern Recognition and Graphics; Python; Software Engineering/Programming and Operating Systems; Source code; the New AI Revolution; Vision; Cluster analysis; Evolutionary computing; Clustering; Framework; Python
• ISSN: 2662-995X; 2661-8907
• DOI: 10.1007/s42979-021-00511-0

EvoCluster is an open source and cross-platform framework implemented in Python language, which includes the most well-known and recent nature-inspired metaheuristic optimizers that are customized to perform partitional clustering tasks. This paper is an extension to the existing EvoCluster framework in which it includes different distance measures for the objective function, different techniques of detecting the k value, and a user option to consider either supervised or unsupervised datasets. The current implementation of the framework includes ten metaheuristic optimizers, thirty datasets, five objective functions, twelve evaluation measures, more than twenty distance measures, and ten different ways for detecting the k value. The source code of EvoCluster is publicly available at http://evo-ml.com/evocluster/ .