PageRank algorithm based on distance attenuation

This paper addresses the problem of identifying important nodes in the network and proposes a distance-decay PageRank (DDPR) algorithm. By introducing a Distance decay factor into the traditional PageRank model, the influence of nodes on neighbors weakens as the distance of the shortest path increases. It is more in line with the information dissemination laws of the actual network. To verify its effectiveness, this paper combines the personalized PageRank (PPR) algorithm and the SI propagation model to conduct experiments on six real datasets, and analyzes the differences of the optimal parameters under different network efficiencies. The results show that DDPR tends to have larger attenuation parameters in high-efficiency networks to highlight the contribution of nearest neighbors, while in low-efficiency networks, it tends to have smaller parameters to retain the influence over long distances. Compared with standard PageRank and PPR, DDPR performs better in key node identification and dissemination effect, proving that it has high practical value in actual network analysis.