ChatTrack: Chat Room Topic Detection Using Classification

The traditional analysis of Internet chat room discussions places a resource burden on the intelligence community because of the time required to monitor thousands of continuous chat sessions. Chat rooms are used to discuss virtually any subject, including computer hacking and bomb making, creating a virtual sanctuary for criminals to collaborate. Given the upsurge of interest in homeland security issues, we have developed a text classification system that creates a concept-based profile that represents a summary of the topics discussed in a chat room or by an individual participant. We then discuss this basic chat profiling system and demonstrate the ability to selectively augment the standard concept database with new concepts of significance to an agent. Finally, we show how an investigator can, once alerted to a user or session of interest via the profile, retrieve details about the chat session through our chat archiving and search system.