Evolutionary computing for unsupervised clustering methods
Do, Anh Duc
Date of Issue2009
School of Computer Engineering
Clustering represents a core research area of machine learning. It has been widely used in data processing and system learning where characteristics of the feature vectors, such as "localization" are defined or learned. Clustering algorithm attempts to organize unlabeled feature vectors into clusters such that within the same group, feature vectors are considered to be more similar than others of different groups. Among available clustering methods, Hard C-means (HCM) clustering represents non-overlapping clustering category while Fuzzy C-means clustering (FCM) represents the overlapping category. FCM enhances HCM with the introduction of fuzzy concept which is deemed closer to human cognition system.
DRNTU::Engineering::Computer science and engineering
Nanyang Technological University