A clustering method of WAP log

This paper presents a clustering method based on genetic algorithm (GA) for Wireless Application Protocol (WAP) log. Firstly WAP log files are pre-processed and the behavior features of WAP service users are extracted. Then the similarity matrix is constructed to compute the similarities between the behavior feature vectors of WAP service users. Furthermore, compute near-optimal solutions by mapping relationships among vectors into two-dimensional plane and optimizing iteratively with GA so that Euclidian distances among patterns converge to their similarity degrees. The proposed clustering method of WAP log is flexible and fast. Experimental results show the method is effective and feasible.