A Web Service Implementation for Large-Scale Automation, Visualization, and Real-Time Program-Awareness Via Lexical Link Analysis

Report Number: NPS-AM-11-186

Series: Acquisition Management

Category: Data Analysis

Report Series: Sponsored Report

Authors: Ying Zhao, Shelley P. Gallup, Douglas J. MacKinnon

Title: A Web Service Implementation for Large-Scale Automation, Visualization, and Real-Time Program-Awareness Via Lexical Link Analysis

Published: 2011-09-01

Sponsored By: Acquisition Research Program

Status: Published--Unlimited Distribution

Research Type: NPS Faculty

Full Text URL: http://acquisitionresearch.net/files/FY2011/NPS-AM-11-186.pdf

Keywords: Lexical Link Analysis, text mining, data mining, Program Elements, Major DoD Acquisition Programs, Universal Joint Task Lists, resource

Abstract:

DoD acquisition is an extremely complex system, comprised of myriad stakeholders, processes, people, activities, and organizational structures. Processes within this complex system are encumbered by the continuous creation of large amounts of unstructured and unformatted acquisition program data, which is narrowly useful, yet difficult to aggregate across the enterprise. Acquisition analysts and decision-makers must analyze this available data to obtain a complete and understandable picture. This is a kind of systems non-congruence which has been difficult to overcome. For those embedded within the complexities of the acquisition community, this effort represents a daunting, if not impossible, task. We will apply a data-driven automation system, namely, Lexical Link Analysis (LLA), to facilitate acquisition researchers and decision-makers to recognize important connections (concepts) that form patterns derived from dynamic, ongoing data collection. The LLA technology and methodology is used to uncover and display relationships among competing programs and Navy-driven requirements. In the past year, we tested our method using samples of acquisition data for validity. LLA was demonstrated to discover statistically significant correlations, and automatically extract the links that might require expensive manpower to perform otherwise. This year, we started to develop LLA from a demonstration to an operational capability and facilitate a wider range of acquisition research applications. The resulting methodology can facilitate real-time awareness, reduce the workload of decision-makers, and make a profound impact on the long term success of acquisition strategies by revealing the current status of acquisition programs, and connections within and external to contributing or competing interests, as well as inform potential strategic choices available to decision-makers.