Identifying and tracking online financial services through web mining and latent semantic indexing

As Internet usage has heavily increased within recent years, money launderers have started to take advantage of Online Financial Transaction (OFT) services to facilitate their money laundering activities. However, law enforcement has struggled to understand and detect OFT services that criminals use for money laundering. To assist law enforcement in its efforts to identify and monitor OFT services, we have designed the Online Financial Transaction Services Identification Tool (OFTSIT), which crawls the Internet and determines the probability that they are OFT services. OFTSIT analyzes a website's content and extracts textual features using latent semantic indexing (LSI). LSI is a text mining approach that can extract a small number (<; 10) of features from more than 40,000 possible words on a website. OFTSIT inputs the LSI discovered features into a generalized linear model to produce the probability that a website is an OFT service. Testing showed that OFTSIT outperforms current method of manual searching. This paper describes the system architecture, algorithms employed to classify OFT services from other websites, and performance testing to demonstrate OFTSIT's operational relevance.