MetaCluster: An open-source Python library for metaheuristic-based clustering problems

Abstract

Clustering, based on metaheuristic algorithms, is a rapidly developing field. Its goal is to use these methods to reframe clustering issues as optimization problems. In this study, we propose an open-source library named MetaCluster. This library leverages the latest metaheuristic algorithms to tackle partitional clustering challenges. MetaCluster has two primary goals: to be open-source and to provide a user-friendly interface with comprehensive documentation. This design aims to facilitate usage for individuals with varying experience levels across diverse applications and domains. The current version of MetaCluster is easy to install and deploy. It offers access to 48 available datasets, over 40 performance metrics, more than 40 evaluation measurement methods, and a wide array of 200 metaheuristic algorithms.

Keywords

• Metaheuristic algorithms
• Clustering
• Evolutionary computation
• Optimization