Interpreting Graph-Based Sybil Detection Methods as Low-Pass Filtering

Online social networks (OSNs) are threatened by Sybil attacks, which create fake accounts (also called Sybils) on OSNs and use them for various malicious activities. Therefore, detection is a fundamental task OSN security. Most existing methods based the graph structure of OSNs, have been proposed recently. However, although almost all compared experimentally in terms performance noise robustness, theoretical understanding still lacking. In this study, we show that graph-based can be interpreted unified framework low-pass filtering. This enables us to theoretically compare analyze each method from two perspectives: filter kernel properties spectrum shift matrices. Our analysis reveals depends effectiveness Furthermore, basis analysis, propose novel SybilHeat. Numerical experiments synthetic graphs real demonstrate SybilHeat performs consistently well with structural properties. study lays foundation leads better methods.