Attack Detection in Recommender Systems

Unsupervised and supervised methods for the creation and detection of hurriedly created profiles in recommender systems

This code is an implementation of attack detection methods proposed in [1].
We propose a framework to create and identify anomalous rating profiles, where each attacker (outlier) hurriedly creates profiles that inject into the system an unspecified combination of random ratings and specific ratings, without any prior knowledge of the existing ratings.

You can find more details in https://sites.google.com/site/costaspanagiotakis/research/hurryattackrs

Usage 1: runDetection.m : Run the detection methods proposed in [1]
Usage 2: createDatasetsWithAttack.m : Creates datasets with attacks proposed [1]

In the folder Attacks you can add the datasets from
https://sites.google.com/site/costaspanagiotakis/research/hurryattackrs
or you can run createDatasetsWithAttack. This folder is used by runDetection.m

In the folder origDatasets you can add the three original datasets (ml.txt, ml100k.txt and sn.txt) from
https://sites.google.com/site/costaspanagiotakis/research/hurryattackrs
this folder is used by createDatasetsWithAttack.m

The folder AttacksTrainVal is used in the code to write some results

We will appreciate if you cite our paper in your work.

[1] C. Panagiotakis, H. Papadakis, and P. Fragopoulou, Unsupervised and Supervised Methods for the Detection of Hurriedly Created Profiles in Recommender Systems, International Journal of Machine Learning and Cybernetics, 2020.

[2] C. Panagiotakis, H. Papadakis and P. Fragopoulou, Detection of Hurriedly Created Abnormal Profiles in Recommender Systems, International Conference on Intelligent Systems, 2018.