Intelligent Classification and Clustering Techniques for Text Data Mining

For Federal agencies' management programs, there exists aspecific need for an integrated software analysis suite which can:(1) process online information relevant to their needs,(2) provide pattern and trend identification, and(3) link solicitations andrequirements documents to open-source suppliers,research and development capabilities.Automated information retrieval and document classification hasbecome one of the most important technologies for web-basedapplications.Integration of data miningalgorithms with textual analysissystems, termed Text Data Mining (TDM), represents a promisingapproach to such a knowledge management problem.The objective of this Phase I research is todesign an overall system architecturefor an InformationIntelligence-based Program Management System, andinvestigate classification and clustering techniquesto analyze document collections, classify incoming documentsand identify trends within the subject areas.In particular, we will investigate the use ofselected statistical,Artificial Intelligence (AI) and Neural Networks (NN) techniques forimproving the classification and clustering performance of TDM systems.Profesors Daniella Rus (Dartmouth College) and Joydeep Ghosh (University of Texas)will provideconsulting support for this Phase I.The Phase I base will investigate the feasibility of the proposedapproaches for TDM. The Phase I option will create a detaileddesign specification for a prototype IIPMS.Commercial applications of the proposed technologyinclude all private sector companies, federal and state agencies with either technology requirements, or products and services for sale.It will appeal to high-technology businesses, prime contractors,small/medium companies,individual consultants,innovators, university and federalresearch institutions as a cost-effective alternative to traditional(in-house) sales and marketing.Federal, state and local agencies, prime contractors andother businesses are increasingly in favor of acquisition andlicensing oftechnology items instead of in-house engineering (buy vs. build).The total potential market for the company's products and services isestimated to be over one million subscribers worldwide.