Personalized Recommendation System on Massive Content Processing Using Improved MFNN

摘要

Though the research in personalized recommendation systems has become widespread for recent years, IEEE Internet Computing points out that current system can not meet the real large-scale e-commerce demands, and has some weakness such as low precision and slow reaction. We have proposed a structure of personalized recommendation system based on case intelligence, which originates from human experience learning, and can facilitate to integrate various artificial intelligence components. Addressing on user case retrieval problem, the paper uses constructive and understandable multi-layer feedforward neural networks (MFNN), and employs covering algorithm to decrease the complexity of ANN algorithm. Testing from the two different domains, our experimental results indicate that the integrated method is feasible for the processing of vast and high dimensional data, and can improve the recommendation quality and support the users effectively. The paper finally signifies that the better performance mainly comes from the reliable constructing MFNN.