An Integrated Suite of Text and Data Mining Tools for Program Managers

"This proposal describes an effort to build an integrated suite of tools incorporating text mining and data mining tools for information extraction and knowledge discovery from requirement sources and text databases. Successful management depends in part onidentifying and understanding requirements, discerning linkages among requirements (e.g., commonality, dependency, priority, etc.), and recognizing correspondence between program requirements and the capabilities of available resources. Requirements takeseveral forms, but of particular interest are large written documents, such as Software Requirements Documents, Strategic Plans, and R&D Master Plans. Requirements may also originate from databases of similar software projects, operating experience andmaintenance information. In either the database form or the resulting documents, mastery of these information sources presents a daunting challenge. The technologies of text and data mining have great potential for assisting Managers in their task ofdefining or understanding requirements from these very large data sources by identifying relationships among requirements and discovering connections between the requirements and other R&D activities reported in bibliographic databases. In Phase 2 we willbuild a software toolset that will leverage the existing Tech OASIS text-mining tool and extend it with enhancements for mining large requirements documents and cross-mining the results