Recommendation system using a deep learning and graph analysis approach

When a user connects to the Internet fulfill his needs, he often encounters huge amount of related information. Recommender systems are techniques for massively filtering information and offering items that users find them satisfying interesting. The advances in machine learning methods, especially deep learning, have led great achievements recommender systems, although these still suffer from challenges such as cold-start sparsity problems. To solve problems, context communication network is usually used. In this article, we proposed novel recommendation method based on matrix factorization graph analysis namely Louvain community detection HITS finding most important node within trust network. addition, leverage autoencoders initialize latent factors, Node2vec embedding gathers users' factors graph. implemented Ciao Epinions standard datasets. experimental results comparisons demonstrate approach superior existing state-of-the-art methods. Our outperforms other comparative methods achieves improvements, is, 15.56% RMSE improvement 18.41% Ciao.