Evaluating and Predicting Contract Performance Using Machine Learning: A Feasibility Study
Abstract
The Office of Acquisition Analytics and Policy, which is part of Office of the Undersecretary of Defense for Acquisition and Sustainment (OUSD(A and S)), tasked the Institute for Defense Analyses with assessing the feasibility of using text analytics and machine learning to evaluate contracts and predict program performance. Although the machine-learning algorithms we evaluated could not predict program performance using the metrics we applied to the contracts, we were able to use text analytics to extract useful information from these contracts and similar documents. This report describes machine-learning algorithms that are feasible for exploratory and descriptive contract analysis. However, further research is necessary to establish whether it is reasonable to use this information, with additional data sources, to predict program performance.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2020
- Accession Number
- AD1121596
Entities
People
- Brian G. Gladstone
- Gregory A. Davis
- Laura A. Hildreth
- Miranda G. Seitz-mcleese
- Travis L. Depriest
Organizations
- Institute for Defense Analyses