Hybrid recommendation approach enhanced by deep learning

Collaborative filtering based on matrix factorization has been very successful,while cold-start and data sparseness problems have not been well resolved.Hence,many studies have attempted to include review information into rating predictions.This paper presents a hybrid model that introduces deep learning into recommendation system with collaborative filtering.The algorithm combines a stacked denoising auto encoder (SDAE) with a latent factor model (LFM) to make use of both review and rating information to improve the rating predictions.Evaluations on a large,commonly used Amazon dataset show that this approach significantly improves the rating prediction accuracy in comparison with traditional models,with up to 64.43％ better predictions.