Abstract:

A novel two-level multiple neural networks-based collaborative filtering recommendation algorithm (TMNN-CFA) for rating prediction is presented. By cooperating the multiple back propagation (BP) networks together, the higher layer neural network propagates conversely the output deviation until to the lower layer neural networks to amend the network weights and based on which, item recommendation is accomplished in the forward process of two layers networks relying on the factors such as ratings, etc. Experiment results on the standard Movielens show that TMNN-CFA method is effective and feasible for item recommendation.

Key words: back propagation neural networks, item recommendation, collaborative filtering