A novel collaborative filtering framework based on Fuzzy C-Means clustering using FP-Tree

Collaborative filtering (CF) has been a comprehensive approach in recommendation system. But data are always sparse; any given user has seen or buys only a small fraction of all items. This becomes the bottleneck of CF. Cluster-based smoothing technique for nature language processing is successful to estimate probability of the unseen term by using the topic (cluster) of the term belongs to, which motivate us to examine the sparsity problem on collaborative filtering. There are many clustering algorithms. Many K-Means-like algorithms which have low time and space complexity have been widely used. But the E-Commerce data are always mixed-type, like numerical, nominal, set etc. The K-Means-like algorithm is not suit to handle these mixed-type. In this paper, we use FP-Tree successfully to handle nominal and set type in mixed-type data. We chose an improved Fuzzy C-Means algorithm which leverages a FP-Tree approach to cluster the items. As a result, we provide higher accuracy as well as increased efficiency in recommendations. The experiments show that improved FCM algorithm clusters items efficiently and raises the efficiency of recommendations.