Home

Awesome

Spammers on social network detection

<p> <center> (NTDS EPFL, Fall 2018)</center> </p> <p> Feature extraction, classification , and prediction of spammers on social network

In this project we focus on detecting spammers in a social network. The ultimate purpose of this project to identify (i.e., classify) the spammer users based on their relational and non-relational features.</p> See the report for more details. </p>

Running the code

The dataset represent users from tagged.com social network. The full dataset is provided in : https://linqs-data.soe.ucsc.edu/public/social_spammer/. </p>

Please think to change the name of the csv files to be able to run the code. These are the name of the different datasets we worked on :

These filtered relations files should be generated using the spammer_subnetwork.ipynb after downloading the full dataset from the link above. </p>

The repository contains four jupyter notebooks :

More details

See the project report provided also in the repository. The slides presented is also provided.

Authors

<p> <center> (©) EPFL-NTDS-Group20-Fendri Hedi,Jeha Paul,Nguyen Minh Nguyet,Mantonanaki Christina </center> </p>