The Research And Application Of Hybrid Recommendation Algorithm In The Cloud Computing Platform

With the development of Internet and the growth of information, information overload is becoming more and more serious. Because of the complexity of Internet, how to extract the information to meet the needs of users become urgent problems. Recommendation system is born in such a background. Recommendation system can dig up information effectively to meet the needs of users according to a certain algorithm. With the development of the recommendation system, recommendation algorithm also emerge in endlessly. So this article choose recommendation algorithm as one of the focus of research.Collaborative filtering is one of the most mature recommendation technology. Due to the score of extremely sparse, the quality of recommendation drops. This paper proposes a hybrid recommendation algorithm in combination with the project properties to solve sparse matrix problems. Hybrid algorithm combine the similarity of Item-Based CF and project properties to improve the accuracy of the prediction. Experiments shows that the proposed hybrid recommendation algorithm has better prediction performance than the algorithm of collaborative filtering recommendation.Although hybrid recommendation algorithm has better accuracy, the operation efficiency of the algorithm are also decreases. So this article combined recommend algorithm with cloud computing. This paper uses the Hadoop platform. Through the MapReduce parallel process of the calculation, improves the computational efficiency of the algorithm. The experiments show that hybrid recommendation run on Hadoop cloud computing platform can has better parallel performance, and also proves the advantage of highlights the Hadoop cloud computing platform.The article has a deep research in recommendation algorithm and Hadoop. Aiming at the problem of collaborative filtering, the article propose a Hybrid recommendation algorithm that run on the Hadoop platform to improve the operation efficiency. Experiments show that hybrid recommendation running on the Hadoop has a better quality and higher efficiency.