A New Approach to Robust k-Means Clustering Based on Fuzzy Principal Component Analysis

摘要

PCA-guided k-Means performs non-hierarchical hard clustering based on PCA-guided subspace learning mechanism in a batch process. Sequential Fuzzy Cluster Extraction (SFCE) is a procedure for analytically extracting fuzzy clusters one by one, and is useful for ignoring noise samples. This paper considers a hybrid concept of the two clustering approaches and proposes a new robust k-Means algorithm that is based on a fuzzy PCA-guided clustering procedure. In the proposed method, a responsibility weight of each sample in k-Means process is estimated based on the noise fuzzy clustering mechanism, and cluster membership indicators in k-Means process are derived as fuzzy principal components considering the responsibility weights in fuzzy PCA.