Lexical Link Analysis Application: Improving Web Service to Acquisition Visibility Portal

Abstract

Lexical Link Analysis (LLA) is a form of text mining in which word meanings are represented in lexical terms (e.g., word pairs) of a word network. In the past, we have shown how LLA can systematically and automatically discover new patterns in large-scale defense acquisition data of multiple programs as indicators for program or investment performances. We also started to apply LLA to understand the quality of the data by comparing categories of information, detecting data and gaps. Last year, we examined the Acquisition Visibility Portal (AVP), which is a critical tool that provides the DoD-wide acquisition community with authoritative and accurate data services. We reported the first program from AVP to have undergone a relatively comprehensive LLA analysis. This year, we found that there is much consensus or consistency in the various categories (e.g., acquisition and engineering communities) of artifacts, yet gaps or low correlations seem to characterize the majority of the examined data for the relations among these categories. LLA, however, is able to discover in detail where the gaps and inconsistencies of the data reside. The depicted findings offered in this report can help decision-makers improve their resource and big data management, to better understand how particular acquisition strategies may affect the desired return on investment (ROI) among projects.