Manipulability of PageRank under Sybil Strategies
Alice Cheng, Eric Friedman · 2006
The sybil attack is one of the easiest and most common methods of manipulating reputation systems. In this paper, we quantify the increase in reputation due to creating sybils under the PageRank algorithm. We compute explicit bounds for the possible PageRank value increase, and we use these bounds to estimate the rank increase. Finally, we measure the effect of sybil creation on nodes in a web subgraph. We find that the resulting rank and value increases agree closely with the analytic values.