bsnsk / deanonymization

deanonymization

De-anonymization

By Shuyang S (and Jialin L, Yuemei Z) | Started @ Sept, 2015

---

Contents

• Project Description
• Skeleton
• Performance
• Experiment Tool

Description

Social network de-anonymization (provided with an anonymized graph and a crawled graph)

Skeleton

The process is divided into two parts: firstly generate a similarity matrix for node pairs acrossing two graphs, and then use specific methods to match them into answer pairs.

Performance

Our algorithm has a satisfying performance with considerable improvement compared with baseline algorithm.

Experiment Tool

Use run.py.

Requirement

• File soc-Livejournal1.txt placed at the root directory of this repository.
• Cmake installed.

Usage

Format: ./run.py [T] [M] [N] [O], where

• T is the number of dataset to be run;
• M is the method of anonymization, where
  • 0 for naive,
  • 1 for sparsify, and
  • 2 for switching;
• N is the number of nodes in the subgraph which is used to generate two graphs;
• O is the number of nodes in the overlap part.

And therefore the number of nodes in generate graphs should be (N - O) / 2 + O.

An example is below:

$ ./run.py 1 0 10000 5000

Find your result at the folder result/.