Evaluation of Monte Carlo Method in PageRank
Zhi Quan Zhou · MOspace Institutional Repository (University of Missouri) · 2015
PageRank is one of the principle ranking algorithms. This method is interpreted as a frequency of visit. As the number of web pages increasing dramatically, the time to finish updating takes longer. Alternatively, Monte Carlo method provides good estimation of the PageRank for relatively important pages after simulating random walks. The Monte Carlo score is calculated as the visited times to the particular page verses the total number of visited pages during the random walks. This estimation can be used as the initial vector for the Power Iteration. This process takes less iteration numbers than the original PageRank to converge. Then the time for the entire ranking procedure is reduced. Using Monte Carlo method together with the Power Iteration in web ranking is feasible.