Researchers Develop Technique to Identify Anonymous E-Mail Authors
Led by Prof. Benjamin Fung, the team developed a method based on pattern identification techniques used in speech recognition and data mining. The first step involves analyzing a suspect's e-mails by identifying patterns in other messages that he or she has sent. Once these patterns are identified, investigators can filter any trends that appear in e-mails sent by other suspects.
The leftover patterns are then scrutinized more closely, with special attention paid to something known as a 'write-print' -- a collection of characteristics that effectively amount to a kind of digital fingerprint. "Let's say the anonymous email contains typos or grammatical mistakes, or is written entirely in lowercase letters," Fung explained. "We use those special characteristics to create a write-print. Using this method, we can even determine with a high degree of accuracy who wrote a given email, and infer the gender, nationality and education level of the author."
Incredibly enough, the researchers found they could identify suspects with about 80 to 90-percent accuracy. According to Fung, this kind of accuracy implies that the method could have very real-world applications. "Our technique was designed to provide credible evidence that can be presented in a court of law," Fung said. "For evidence to be admissible, investigators need to explain how they have reached their conclusions. Our method allows them to do this."