Research Of Book Recommendation System Based On Hadoop

In the field of E-book,there is lots of user's information about books.So excavating readers' reading preferences from so much information and providing personalized recommendation to users are very important for improving user's experience.In addition,with the increasing number of users,books and data,scalability has become a bottleneck restricting the development of recommender system.Putting the application of big data Hadoop platform to recommendation system can solve this problem effectively.Firstly,it analyze recommendation algorithm,fuzzy C means clustering algorithm,describes Hadoop platform's main components and principles,designs the general model and the recommendation system based on a recommendation system architecture,and introduces the design idea of each the modules in the system.Secondly,it puts forward the Slope One algorithm based on Clustering.It does fuzzy clustering to user,and then it finds the Top-k nearest neighbors in nearest neighbor classes and calculates the similarity between them,and it uses this as the weight of the difference between the items.Considering the change of user's interest over time,it uses the improved time attenuation function to process the score value of the project.In order to avoid the blindness of the fuzzy clustering algorithm for the selection of the initial cluster center and the number of clustering,the Canopy algorithm is used to optimize it.The parallel processing of the improved algorithm is realized on the Hadoop platform.Then,it uses the watercress user-book grading dataset to test and analyze the performance of the improved recommendation algorithm in the Hadoop cluster.The improved algorithm is much more accurate than the traditional algorithm seen from the experimental results.In addition,it shows the effectiveness and efficiency of the improved algorithm by comparing the performance of the algorithm under the distributed environment and the stand-alone mode and the higher implementation rate of the proposed algorithm.Finally,the improved algorithm is realized,and the results of the book recommendation are displayed.