The use of textual data mining in the detection of the proliferation of weapons of mass destruction

This dissertation describes an experiment where a form of textual data mining of the World Wide Web is performed to discover web pages that may be indicative of the construction, research about, or at least interest in the various weapons of mass destruction (Biological, Chemical, Nuclear, and Radiological). Words have been selected from various control lists (Australia Group, Military Critical Technologies List, and others) and assigned various point scores according to their importance in constructing a weapon. Web pages from the World Wide Web are crawled, and analyzed and a score in obtained for each page scanned in each of the weapon categories. An innovative computer man-machine interface presents the information to an intelligence analyst in a manner such that the most important sites are analyzed first. The results of the experiment are analyzed and suggestions for future enhancements to the program are made.