The Study Of Deep Learning Based Recommender System Algorithm

With the rapid growth of information technology,more and more information are generated every day,which makes people difficult to find their interested information quickly.To solve this information overload problem,the recommendation system can take an important role.In the recent years,deep learning has achieved great progress in many fields,like the computer vision,human-computer game and natural language processing.Deep learning can get better performance than many traditional machine learning algorithm.People also take a lot research of deep learning in the recommendation system area,and got many achievements.It has proved that deep learning can get better performance and adapt to many scenarios than the traditional recommendation algorithm like the collaborate filter or content based recommendation algorithm.In this paper,I will review the recent deep learning research achievements in recommendation area,and will focus on three deep learning models,the auto encoder,LSTM network,Wide & Deep network.And will discuss how to use these models in the real world recommendation system based on the movielens dataset.