Application of Natural Language Processing to Defense Acquisition Executive Summary Reports

Abstract

Major Defense Acquisition Programs (MDAPs) are required to submit quarterly Defense Acquisition Executive Summary (DAES) reports which, among other information, contain ratings for each program area (green, yellow, red, etc.) and explanations of these ratings by the program manager. Natural language processing, a powerful machine learning tool, can harness the wealth of text data available in these reports in order to predict the ratings given the program manager’s explanation in the report. With this information, the model can be used to indicate which programs are not reporting their ratings as expected in order to indicate which programs may need further investigation. Utilizing machine learning in this manner can increase insights into data in the DAES reports and has broad implications for further applications of these techniques to other acquisition data.