Study Of Recommendation Algorithms Based On Deep Learning

With rapid development of Internet technologies,a great variety commodities show up.It makes users can not pick up goods quickly which are satisfying or fit for them through viewing the description of goods by themselves,in order to enhance purchasing efficiency,recommendation algorithms are born at the right moment.They are a kind of machine learning algorithms which connect to human life closely and they can not only bring convenience to the users but also deliver huge benefits for merchants,therefore,they can be called “win-win”algorithms.Deep learning is a research field which have attracted extensive attention of scholars.The deep architecture make them can learn more complicated structure,thus they gain great achievement in lots of research fields,such as speech recognition,machine translation,image recognition.In this article,we combine deep learning and recommendation algorithms together.We apply gated recurrent unit in recommendation algorithms and compare it with other popular algorithms in recommendation filed on two data sets,foreign movie rating data-MovieLens data and domestic movie rating data-douban movie rating data which is crawled by ourselves,respectively.