CSAFE aims to be at the forefront of developing a statistical framework to solve key challenges in digital evidence analysis. In January 2018, CSAFE in partnership with NIST initiated two new research projects to address critical needs in digital crime.
New research area – Digital Crime Scene Reconstruction and Cyber Bullying
In today’s society, social activities frequently take place in an online setting. New social networks such as Twitter, Facebook, emails and more are becoming more prone to harassing behaviors. Potential reasons for the increased incidence of cyber bullying and harassment include a perceived sense of anonymity and perceived lack of social norm pressures due to lack face-to-face interactions.
CSAFE researcher and Carnegie Mellon associate professor in machine learning Pradeep Ravikumar will lead a team of investigators using statistical machine learning to provide narrative reconstruction of digital crime scene events. Researchers will focus on answering the “what, the how, and the why” of cybercrimes with an emphasis on online bullying and harassment.
Using a machine learning system based on training data from available datasets, researchers will evaluate how much a piece of evidence can be trusted and infer both how and why pieces of digital evidence were created.
New research area – Matching Seller Accounts on Online Anonymous Marketplaces
William F. Eddy, CSAFE Co-Director and John C. Warner professor of statistics, emeritus at Carnegie Mellon University will co-lead a team with graduate student Xiao Hui Tai to investigate anonymous marketplaces on the dark web. These marketplaces have been in the spotlight due to the sale of illicit products, particularly drugs.
The new CSAFE project aims to link seller accounts on anonymous marketplaces by analyzing various forms of data to generate predicted match probabilities. Information on unique sellers will be used for further analysis, such has estimating seller volumes, tying seller accounts to particular countries to track sales and more.
The CSAFE approach is the first known attempt at using statistical matching techniques for marketplace seller data and identifying deceptive behavior.
Expanding the CSAFE Research Portfolio
These additional projects add to CSAFE’s mission of providing law enforcement and forensic examiners with statistically based techniques for digital evidence analysis. As CSAFE continues to expand its expertise in digital forensics, we aim to increase insight and conclusions for digital evidence analysis.
Visit the CSAFE blog to learn more about how digital evidence changes the world of crime, and read about another CSAFE digital evidence research project on steganalysis in the news section.