Peer-to-peer file sharing networks such as Gnutella and BitTorrent pose a unique challenge to law enforcement officials. As well as facilitating the illegal distribution of copyrighted material such as games, films and music, such networks are being used for the purposes of serious organised crime, including the distribution of child abuse media and other illegal pornographic material
As a response, we have conducted an exploratory study into the application of collaborative filtering techniques in peer-to-peer network forensics, looking at applications such as media identification, downloading behaviour prediction and distance metrics between peers. We find promising results for the application of collaborative filtering techniques to media identification and make initial validation of its use as a distance metric.
Below are a selection of relevant documents and results, including R script files relating to the analysis of the results. Note that some intermediate processing was carried out on results in Bash and R so inputs are not necessarily aligned for immediate repeatability.