Sparse Projected Maximum Entropy Fuzzy c-Means

Clustering approaches have been widely used for discovering unknown features in massive data. Fuzzy clustering is an alternative to conventional hard clustering which assigns objects to more than one cluster. In recent decades, dimensionality reduction clustering has attracted attention because of the demand for interpretability when dealing with high-dimensional data. In this paper, we propose a novel approach called Sparse Projected Maximum Entropy Fuzzy c-means (SPEFCM). SPE-FCM is a combination of sparse dimension reduction and maximum entropy fuzzy c-means. The benefits of our approach are that it can correctly extract the features related to the true cluster structures and the clustering results can be interpreted in low-dimensional space. In this paper, comparative experiments are conducted on synthetic and real-world datasets to verify the performance of the proposed approach. The experimental results show that SPEFCM is more competitive than comparative approaches.