Using Neighborhood Information for Non-image Data via Fuzzy Clustering

Abstract

We provide an improved version of the classic Fuzzy C-Means (FCM) technique designed to take use of neighborhood 
information in Euclidean-space non-image datasets. Our innovative approach seeks to enhance the intrinsic fuzziness of individual 
data points by using spatial contextual clues present in such datasets. Our method goes beyond the constraints of traditional FCM 
by using neighborhood information, which improves clustering performance. We demonstrate our method's advantage over 
traditional FCM in capturing spatial connections within the data by validating its effectiveness on both synthetic and actual 
datasets. Our results demonstrate how well our method works to improve clustering results by carefully integrating neighborhood 
information into the FCM framework for non-image data in Euclidean space.