Year
2007
Abstract
As part of its mandate, the Canadian Safeguards Support Program (CSSP) is responsible for providing assistance to the IAEA in areas of knowledge management. One of the tools that the CSSP has supplied to the IAEA is VITA, a Visual Interface for Text Analysis. VITA is a meta-search engine developed by the CSSP in cooperation with Health Canada and the Department of National Defense. VITA can aggregate the results from many search tools into a single 3-D visual representation of the found documents. One application is to facilitate the process of uncovering clandestine nuclear supply networks. VITA allows an analyst to locate, organize, summarize, and present documents of potential interest as a 3-D picture with object size and color indicating document relevance. VITA visually groups like documents together, thus enabling quick refinement of the search process. The 3-D representation can be rotated in space to observe and investigate the relationships (often unknown and unexpected) among documents. The software isolates and calls attention to otherwise unsuspected documents of importance in ways that other search tools cannot. The latest version of VITA adds clustering to its already powerful analytical features. Clustering reveals hidden relationships within the found documents – relationships that where not evident in the search terms. This paper explores how this clustering process works and demonstrates some of the (sometimes startling) information it reveals. VITA improves the quality and speed at which intelligence analysts can explore massive text corpora.