Abstract
With the rapid development of uncertain and large-scale datasets, Fuzzy Possibilistic C-means Clustering (FPCM) and Granular Computing (GrC) were introduced together with the aim to solve the feature selection and outlier detection problems. Utilizing the advantages of the FPCM and GrC, an Advanced Fuzzy Possibilistic C-means Clustering based on Granular Computing (GrFPCM) was proposed to select features as a preprocessing step for clustering problems and granular space is used to handle the uncertainty. Experimental results reported for various datasets in comparison with other approaches exhibit the advantages of the proposed method.