Finding Semantically Related Terms on the Web Using Random Indexing : Detecting New Drugs on the Dark Web

Abstract: There is an emerging international phenomenon of new drugs that have not previously existed being manufactured and then traded over the Internet. In many cases law enforcement agencies and customs have problems keeping track of all these new drugs causing them to have problem to make appropriate responses to the situation. Because of this measures are now taken to discover these new drugs early as they emerge by developing systems raising warnings when they recognize patterns of new drug trends. This project investigates ways to find these new drugs as they appear and are being discussed on the Web. Techniques able to interpret the semantic meaning of words, such as the technique random indexing, which has the ability to describe the meaning of words in vectors, are investigated and used to process text gathered from appropriate Internet sources. It is shown that this technique has the capacity to find and extract drug names from text and could be an appropriate component of future systems discovering new drugs being discussed on the Internet.