An Evolutionary Fuzzy System for Multiclustering in Data Streaming

Abstract

The rapid advance of technology has led to exponential growth in the generation of data in various fields, revolutionising how information is captured, processed and analysed. The continuous flow of data has proven to be a valuable resource for agile decision-making. The complexity of being able to make decisions from different sources has led to the emergence of new techniques that aim to obtain as much information as possible to relate data efficiently, taking into account multiple perspectives.A new approach to clustering streaming data from different perspectives is introduced in this paper. It is based on evolutionary algorithms (CHC), fuzzy rules and consensus function. The proposed approach is called FuzzyMultiCHCClust-DS. This approach addresses the clustering of streaming data by optimising the clustering process taking into account traditional clustering perspectives such as k-means, Clara, Pam, Fanny and MiniBatchKMeans. It has been tested using different sets of linguistic labels, achieving a high degree of compactness and separation of the clustered data.