Research And Realization Of University Library Books Recommendation System Based On Hadoop

In recent years,personalized recommendation service has been applied to many areas of society,typically like e-commerce,news,video and so on.However,at present,the application of personalized recommendation service in the library is still in its infancy and many university libraries do not provide personalized recommendation services.At present,there are a large number of books in the library of universities.It becomes more and more difficult for the readers to discover the books which they are interested in through the traditional search engine.This also poses actual requirements for the research on the books recommendation system of the university library.Since the current single recommendation algorithm has its own advantages and disadvantages on the recommendation effect,this paper proposes to apply the hybrid recommendation strategy to the book recommendation system in order to achieve better recommendation results.Furthermore,the collection of books in university libraries almost exponentially increases each year.The traditional recommendation algorithm based on stand-alone mode has serious performance bottlenecks in the face of massive data,which is hard to meet the actual computing needs.However,the parallel computing framework based on hadoop provides a new solution to this problem.The main contents of this paper are as follows:(1)By analyzing and studying the current mainstream recommendation algorithms(collaborative filtering recommendation algorithm based on users,collaborative filtering recommendation algorithm based on articles,bipartite graph recommendation algorithm based on random walk and recommendation algorithm based SVD),We use the dominant matrix method to give different weights to the four hybrid recommendation algorithms and combine them into a hybrid recommendation strategy.The final experimental results show that the proposed hybrid algorithm has good accuracy and diversity.(2)Through the study of the execution flow and operation principle of MapReduce distributed computing framework,the proposed algorithm adopted in this paper has been implemented in parallel under the hadoop framework.The experimental results show that the proposed algorithm based on parallel computing framework shows great advantages in performance.Based on the research of hybrid recommendation strategy and Hadoop parallel computing framework,this paper designs and implements Hadoop-based university library recommendation system.