Research And Implementation Of Online Literature Recommendation System Based On Citation Network

In recent years, the rapid growth of academic literature, there are more than one million papers published in 2015, the vast amounts of literature brought many reference resources to researchers, but from the mass literature we can find the useful paper in short time has become an urgent problem. It offer a document or retrieve a specific keyword for online literature recommendation system, and then the system can be by automated real-time way to recommend appropriate citations to the user.Faced with this problems, we designed a new online literature recommendation system, the system provides a variety of document formats seed input interface, we can based on user input, by Baiduscholar to build a complete network for references, and then the network by MCL cluster and statistical citations counts by PageRank method to make sure the paper is important. At last we dig out the first layer of the core document, and then we can see the first layer of core document as second seed paper to find second layer of core paper, and find the latest citation network, and then, a complete network has been finished, system ultimately to visually display the results of the literature recommended to facilitate academic research.The system is consists of four modules, first is reference list extraction module, our main purpose to finish document list separation; second is reference network data acquisition module, our main purpose to build Citation Network, third is Citation Network data mining module, we can in the existing network to identify the core literature in the field by data mining methods; the last module is the literature visualization module, using Echarts and vis.js to show recommended literature results.Finally, according to the recommendation literature results we can make conclusions:the system is able to find the core of the literature in the mass of the Citation Network,and with respect to the traditional academic recommendation, compare traditional citation recommendation system to our system, we can also dig out some of the recent years classic paper, and our system has more obvious advantages.