A personalized recommendation model based on Self-Organizing Feature Maps

Personalized Recommendation System has been widely used to help user to deal with information overload, and clustering of customers is the basis to produce the recommendation. Clustering of customers in electronic commerce has distinct characteristics compared with other applications, such as the extreme sparsity of dataset, the frequent alteration of clustering of customers because of transference of user's preference. So, traditional methods work poor in the situation. To address these issues, a novel neural network model is proposed at the basis of Self-Organizing Feature Maps. The model has following features: (1) A Restraint Function is introduced into the basic model of SOFM to solve clustering of sparse data. (2) Splitting process and merging process of neurons are constructed to realize dynamic clustering. Experiments show the proposed model has higher capability than general collaborative filtering algorithm.