Web recommendation based on back propagation neural networks

Abstract

Back Propagation Neural Networks are simple, effective, robust and great to estimate predication model, which can be used to combine the latent and extenal features of users and web page for web recommendation.This paper propose a unified collaborative filtering model to get more accuracy recommendation. To discover user communities and prototypical interest profiles, Probabilistic Latent Semantic Analysis Model is applied.The experiment result shows that the technique are the higher accuracy, constant time prediction, and an explicit and compact model representation in the users' Web access log and users' basic information respectively. © 2011 Springer-Verlag.