Fuzzy Clustering Algorithm And Its Application In Collaborative Filtering Recommendation

In recent years, with the rapid amplification of the e-commerce market, vast amounts of product information fill with the life of every customer. In order to let users quickly and conveniently get the required product information, recommendation technology emerges as the times require. As the most successful personalized recommendation technology of recommendation system in present, collaborative filtering recommendation is obtaining more and more attention of researchersThis article describes the common recommendation algorithm and other related technologies. In addition, it compares the various advantages and disadvantages of the recommendation algorithm. It also presents the problems and classification of collaborative filtering recommendation.The article proposes an improvement algorithm which called FCMBMD algorithm. As the scarcity of the sparse data set, the FCMBMD algorithm, which can improve the random selection of the initial cluster centers for fuzzy C-means clustering algorithm and the limitation of distance formula, has been mentioned. Experimental results show that the algorithm regardless correctness, convergence and the number of iterations in the clustering is superior to the traditional fuzzy c-means clustering algorithm.At last, the article proposes an improvement algorithm which based on fuzzy clustering to research collaborative filtering recommendation. Using this algorithm in the user-project scoring model, the large amount of data is plotted to small data set by optimized fuzzy clustering algorithm. So the neighborhood of the user being tested can be got from the small data set, and then get the predictive value of not rated items of aimed user according to the user of neighbor to recommend. The paper mentions that by comparison with experimental results, the fuzzy C-means optimized collaborative filtering recommendation algorithm is better than the FCM-based collaborative filtering recommendation algorithm in clustering and recommended effect.