# walk2friends

**Repository Path**: justyl/walk2friends

## Basic Information

- **Project Name**: walk2friends
- **Description**: walk2friends: Inferring Social Links from Mobility Profiles
- **Primary Language**: Python
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-02-13
- **Last Updated**: 2020-12-20

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# walk2friends
This repository provides a reference implementation of *walk2friends* as described in the paper:<br>
> walk2friends: Inferring Social Links from Mobility Profiles. <br>
> Michael Backes, Mathias Humbert, Jun Pang, and Yang Zhang. <br>
> The 24th ACM SIGSAC Conference on Computer and Communications Security (CCS). <br>
> <Insert paper link>

## Basic Usage
To run the code, please go to folder src/
<br/>``cd src/``

### Attack
Running our social link inference attack on the New York data with each user having at least 20 check-ins:
<br/>``python main_attack.py ny 20``

### Defense
Hiding 60% of the check-ins for defense:
<br/>``python main_hiding.py ny 20 60``

Replacing 60% of the check-ins with a 15 step random walk for defense:
<br/>``python main_replace.py ny 20 60 15``

### Utility
Measuring the utility after hiding 60% of the check-ins:
<br/>``python main_utility_hiding.py ny 20 60``

Measuring the utility after replacing 60% of the check-ins with a 15 step random walk:
<br/>``python main_utility_replace.py ny 20 60 15``

## Requirements
* pandas
* numpy
* scipy
* scikit-learn

It is recommended to install [Anaconda](https://www.continuum.io/downloads), a python data science distribution, which includes all the above packages.

* gensim
* joblib

## Citing
If you find walk2friends useful in your research, please cite the following paper:<br>
> @inproceedings{BHPZ17,<br>
>     author = {Michael Backes and Mathias Humbert and Jun Pang and Yang Zhang},<br>
>     title = {walk2friends: Inferring Social Links from Mobility Profiles.},<br>
>     booktitle = {Proceedings of the 24th ACM SIGSAC Conference on Computer and Communications Security (CCS)},<br>
>     year = {2017},<br>
>     pages = {1943-1957},<br>
>     publisher = {ACM}<br>
>  }<br>

## Miscellaneous
If you have any questions about the code and/or the algorithm, please send an email to <yang.zhang@cispa.saarland>.