Research Of Academic Social Network Modeling And Academic Resource Recommendation Method

In the research work of scientific researchers, publishing papers is an important task. If researchers select an inappropriate publications to submit their new paper, then after a long review process there will be a greater risk of rejection, and the researchers will waste a lot of time and effort. Therefore, for researchers, especially for inexperienced researchers, how to choose an appropriate publication from thousands of journals or conferences to publish their new paper is an urgent problem.This paper studies the technology of academic social network modeling and publication venue recommendation method in the field of academic recommendation to recommend appropriate publications for scientific researchers who need to publish their new papers. Based on the systematic study of co-author network modeling method, personalized recommendation technology based on random walk, and publication venue recommendation technology, we propose a publication venue recommendation method based on co-author network and random walk method. After that, we determine the parameters in our method by experiments, analyze the effect of our method, and compare our method with other existing methods. At last, we deploy our method in a real system to put into use.The publication venue recommendation method based on co-author network and random walk takes co-author relationship between authors and content similarity between the paper to be published and publication venue into consideration. First, we build a co-author network among authors and calculate the co-author relationship weights between authors. Based on the co-author network, we can cluster the authors by using the clustering method based on random walk. Then the author set and paper set related to the paper to be published can be extracted based on the cluster result, after that, the author-paper double layer network can be constructed. At last, we can use the paper content relativity factor combined random walk method to discover the papers which were highly correlated with the target paper, and analyze the publication venues of these papers to recommend appropriate journals or conferences to the scientific researchers of target paper.Experimental results show that our method is better than other contrast methods from the perspective of the success rate, and taking the social relations of authors and content factor of papers into consideration have a significant effect on improving the accuracy of publication venue recommendation and promoting the diversity of recommendation result.