A Chaotic Elephant Herding Optimization-Established Classification (EHOC) for Large Dataset to expand the appearance of Heterogeneous Distributed Situation

Abstract

Data clustering (or classification) is broadly exploited for an enormous range of applications in numerous regions such as education, remedial, invention, and organizations in heterogeneous distributed environments. The large heterogeneous data are processed and examined by using various classification methods to enhance the quality of important information transferring over distributed environment. Here, a Chaotic Elephant Herding Optimization based Classification (CEHOC) is implemented to classify the large amount of heterogeneous datasets to improve the performance of distributed computing. The key motive of CEHOC is to obtain a well arranged distribution of data elements over heterogeneous resources for distributed environment by utilizing some circumstances of chaos theory for population selection. The chaos function is introduced for improving the exploration and exploitation power of search agents in optimization algorithm and for performing the selection of centroids and members of classes (or clusters) optimally and precisely. The MATLAB 2021a tool is used to implement the CEHOC algorithm for four large datasets and the outcomes describe the superior effectiveness of CEHOC algorithm according to parameters like purity index, F-measure, intra-cluster distance, time complexity and standard deviation against previous algorithms such as K-Means, PSO, ACO and EHO.