Web Usage Mining Clustering Using Hybrid FCM with GA
Karunesh Gupta, Manish Shrivastava
Abstract
The most widely used clustering algorithm implementing the fuzzy philosophy is Fuzzy C-Means (FCM) .In this paper, we have proposed a new Hybrid FCM with Genetic Algorithm (GA), we get an improved FCM algorithm which has not only the global search capability of GA but also the local search capability of FCM, and hence can better solve the clustering problem. An improved version of this hybrid clustering algorithm is, therefore, proposed to reduce the number of iterations. The proposed algorithm is tested with MSNBC.com web access logs collected from UCI (University of California Irvine) dataset repository.
Keyword
Fuzzy C-Means, Genetic Algorithm, Fuzzy clustering.