Smarter Government with Text Analytics

27 Jul 2014



Improve Cyber Security with better multilingual search, entity analytics and name matching

Speaker: David Murgatroyd, V.P. of Engineering

Webinar, July 7th, 2014–David Murgatroyd explores how a cyber criminal on the FBI’s Most Wanted list may have been identified and how text analytics could help the government find cyber criminals in a smarter way.

In an opening example, the phrase ‘Chinese Cyber Army’ is explored. Combing the whole of the web for such a phrase is difficult since the most relevant articles are actually in Chinese. Even after translating the query (‘中国电脑部队’) the most important articles will not be found unless smart segmentation is applied (‘中国 电脑 部队‘).

Following the same example, issues surrounding name matching are explored. If investigators wanted to check visa records for 汪东, then they need an English equivalent, because visa applications must be in English. So ‘Wang Dong’? or ‘JackWong’? The answer is not simple and without smart name matching a system will surely identify too many of the wrong people.