Skip to content

Compare Yekta with other anti sybil algorithms

Mohsen Khan-mohammad-zadeh edited this page Aug 23, 2020 · 2 revisions

In all of the algorithms tested on the current sparse graph, Sybils created by attackers can achieve higher scores than the high percentage of honest nodes. The root of the problem is the sparsity of the graph and the fact that most of the honest nodes in the current graph do not have enough degree to achieve more score than the attacker created Sybils.

So an ideal algorithm is the one that

  • Prevents Sybils getting more score than active high degree nodes in the current sparse graph
  • Prevent Sybils getting more score than non of honest nodes when the graph becomes dense and all honest nodes have enough connections

Based on this approach, Yekta is the best algorithm compared to other candidates studied and tested here. You can see comparison results of the Yekta with other algorithms in the both of current sparse graph and a dense graph that 10 dummy connections added to every node to make the graph dense.

Current Sparse Graph

SybilRank

sybil rank

GroupSybilRank

group sybil rank

WeightedSybilRank

weighted sybil rank

LandingProbability

landing probability

NormalizedSybilRank

normalized sybil rank

ClusterRank

cluster rank

SeednessScore

seedness score

Dense Graph

SybilRank

sybil rank

GroupSybilRank

group sybil rank

WeightedSybilRank

weighted sybil rank

LandingProbability

landing probability

NormalizedSybilRank

normalized sybil rank

ClusterRank

cluster rank

SeednessScore

seedness score