Clustering Algorithm Analysis of Web Users with Dissimilarity and SOM Neural Networks

摘要

To effectively organize and analyze massive webinformation, design a web user’s clustering miningalgorithm. SOM neural network algorithm has lots ofdisadvantages, to solve the data clustering, propose a newmethod that uses D-SOM (Dissimilarity-Self Organizingfeature Mapping) algorithm, for clustering web user’s. Thisalgorithm can estimate the center and number of clusteringdata set by dissimilarity computing, optimize SOM neuralnetwork learning and improve clustering effect. Throughdesign the experiment, these web data are collected andprocessed by D-SOM algorithm Experimental results verifywhich D-SOM clustering algorithm has better clusteringaccuracy and imore efficient than SOM neural networkalgorithm.