The study of personalized recommendation based on Web data mining

The coordination filtration technology is one of most successful recommendation technologies at present, but along with the stand structure, the content order of complexity and user population´s increase, coordination filtration technology existence data sparse, extendibility, cold start questions and so on, causes the recommendation effect to reduce greatly, This article proposed the improvement coordination filter algorithm: Based on Web data mining coordination filtration recommendation algorithm, Mainly carries on the recommendation using the connection rule algorithm and the cluster algorithm, focusing on the clustering algorithm as a pre-processing association rules algorithm, Makes a cluster analysis to the data to divide into multiple classes, seeks for the similar user in the kind using the connection rule algorithm, may enhance the recommendation the rate of accuracy, and has carried on the confirmation through the experiment.