Название: Evolutionary Data Clustering: Algorithms and Applications (Algorithms for Intelligent Systems)
Автор: Ibrahim Aljarah, Hossam Faris
Издательство: Springer
Год: 2021
Страниц: 253
Язык: английский
Формат: pdf (true), epub
Размер: 17.1 MB

This book provides an in-depth analysis of the current evolutionary clustering techniques. It discusses the most highly regarded methods for data clustering. The book provides literature reviews about single objective and multi-objective evolutionary clustering algorithms. In addition, the book provides a comprehensive review of the fitness functions and evaluation measures that are used in most of evolutionary clustering algorithms. Furthermore, it provides a conceptual analysis including definition, validation and quality measures, applications, and implementations for data clustering using classical and modern nature-inspired techniques.

It features a range of proven and recent nature-inspired algorithms used to data clustering, including particle swarm optimization, ant colony optimization, grey wolf optimizer, salp swarm algorithm, multi-verse optimizer, Harris hawks optimization, beta-hill climbing optimization. The book also covers applications of evolutionary data clustering in diverse fields such as image segmentation, medical applications, and pavement infrastructure asset management.

Clustering is concerned with splitting a dataset into groups (clusters) that represent the natural homogeneous characteristics of the data. Remarkably, clustering has a crucial role in numerous types of applications. Essentially, the applications include social sciences, biological and medical applications, information retrieval and web search algorithms, pattern recognition, image processing, machine learning, and data mining. Even that clustering is ubiquitous over a variety of areas. However, clustering approaches suffer from several drawbacks. Mainly, they are highly susceptible to clusters’ initial centroids which allows a particular dataset to easily fall within a local optimum. Handling clustering as an optimization problem is deemed an NP-hard optimization problem. However, metaheuristic algorithms are a dominant class of algorithms for solving tough and NP-hard optimization problems.