Adaptive methods for the computation of PageRank Original Research Article

We observe that the convergence patterns of in the PageRank algorithm have a nonuniform distribution. Specifically, many converge to their true PageRank quickly, while relatively few take a much longer time to converge. Furthermore, we observe that these slow-converging are generally those with high PageRank. We use this observation to devise a simple algorithm to speed up the computation of PageRank, in which the PageRank of that have converged are not recomputed at each iteration after convergence. This algorithm, which we call Adaptive PageRank, speeds up the computation of PageRank by nearly 30%.