With the social networks (SNs) becoming ubiquitous and massive, the issue of similarity computation among entities becomes more challenging and draws extensive interests from various research fields. SimRank is a well known similarity measure, howeve