Multicontext Fuzzy Clustering: Toward Interpretable Fuzzy Clustering

Abstract

In this article, fuzzy clustering is employed to establish an innovative clustering approach, aiming to improve and refine the quality of clusters. The development process is derived from the augmented version of fuzzy clustering known as a context-based or conditional fuzzy C-means which efficiently construct linguistic models that preserve interpretability and ability to inference. The objective of this article is to determine data structures under several conditions simultaneously as opposed to a single condition to significantly enhance interpretation feature of fuzzy clustering. The originality of this work is primarily demonstrated by enhancing the quality interpretation of clusters to help in identifying data patterns, and to efficiently reconstruct linguistic models. We developed a rigorous mathematical framework to cluster input space under the influence of several linguistic information granules originated in the output space. The introduced algorithm is quite effective in a vast array of machine learning tasks especially in constructing linguistic models, extracting useful knowledge, and building efficient explainable constructs.