Summary: A System for the Automated Author Attribution of Text and Instant Messages

The paper presents a summary of the ChatSafe system described in “A System for the Automated Author Attribution of Text and Instant Messages” authored by undergraduate students Donais, Peelar, and Roddy with supervisor Dr. Richard A. Frost [1]. ChatSafe is an author attribution system intended for use with short message based communication, i.e. instant messaging or SMS. The authors presented a modified bayesian classifier, used internally by ChatSafe, that improves on the accuracy of a standard bayesian classifier on their metrics [1]. Finally, the authors noted avenues that may be worth pursuing to increase accuracy in ChatSafe. “A System for the Automated Author Attribution of Text and Instant Messages” is the first publication on ChatSafe and is the first work of the undergraduate authors.