doiSerbia

The modified matrix splitting iteration method for computing PageRank problem

Tian Zhaolu (School of Applied Mathematics, Shanxi University of Finance and Economics, Taiyuan, P.R. China)
Liu Xiaoyan (College of Data Science, Taiyuan University of Technology, Taiyuan, P.R. China)
Wang Yudong (College of Textile Engineering, Taiyuan University of Technology, Taiyuan, P.R. China)
Wen P.H. (School of Engineering and Materials Science, Queen Mary University of London, London ENS, UK)

In this paper, based on the iteration methods [3,10], we propose a modified multi-step power-inner-outer (MMPIO) iteration method for solving the PageRank problem. In the MMPIO iteration method, we use the multi-step matrix splitting iterations instead of the power method, and combine with the inner-outer iteration [24]. The convergence of the MMPIO iteration method is analyzed in detail, and some comparison results are also given. Several numerical examples are presented to illustrate the effectiveness of the proposed algorithm.

Keywords: PageRank, Inner-outer iteration, Convergence, Matrix splitting, Power method