AHITS-UPT: A High Quality Academic Resources Recommendation Method

Personalized recommendation technology provides the possibility for users to obtain academic resources quickly and accurately. However, the existing recommendation methods based on user´s historical behaviors and paper contents are limited in terms of expanding user perspectives. The existing methods that evaluate the authority and quality of academic resources based on academic network ignore paper information or consider some unreasonable information, leading to various quality levels of recommendation results. In order to recommend high quality academic resources to users and expand the horizons of users, we propose an Advanced Hyperlink Induced Topic Search (AHITS) algorithm to evaluate the quality and authority of academic resources, propose a user research interest model based on constructing a tripartite graph, namely User-Paper-Topic (UPT), and propose an academic resource recommendation method based on AHITS and UPT. Experimental results show that the methods presented in this paper can effectively remedy the problem that content based recommendation method is not conducive to expand the horizons of users, recommend authoritative authors and high quality papers to users, improve the accuracy of the recommendation results, and effectively reduce the time complexity of algorithm.