Research Of Network Information Analysis Based On Big Data Technology

With the rapid development of Internet technology and communication technology.especially the widespread use of social media,data generated by the network become more and more.In this situation,the recommendation system is widely studied as an effective solution to overload problem.Besides,the mature application of big data technology makes the speed of processing data greatly improved.The research subject of this paper is based on recommendation system technology and its combination with big data technology.This paper analyzes the background of the research and the current research status at home and abroad.After summarizing the domestic and foreign research of recommendation system,this paper mainly studies the neighborhood-based collaborative filtering algorithm and the model-based collaborative filtering algorithm.Collaborative filtering recommendation is the most successful technology,it make nichetargeting recommendation for users according to nearest neighbor rating data.This paper presents a combining collaborative filtering algorithm,which combined KNNitem with RSVD.It took advantage of the local similarity and global similarity,and improved the accuracy of the algorithm.Through experiments which used data from the BYR forum.It proved that improved algorithm has better recommendation effect.In addition,for slow processing speed of recommendation system,this paper do some research of recommendation implementation,then summarized design and implementation method of personalized recommendation system.Finally this paper make construction of a personalized recommendation system make based on the mahout framework.This paper gives a hierarchical design for the personalized recommendation system,introduces the function and principles of each layer,then implements the system through programming.