Data Consolidation of Disparate Procurement Data Sources for Correlated Performance-Based Acquisition Decision Support

Abstract

The study's open architecture framework (i.e., the Cognitive Learning Application Framework [CLAF]) for Acquisition Decision Support and Business Intelligence successfully integrated and prototyped a neural network model using a PMML standard and explored variable relationships using four test hypotheses addressing contract performance data. Regarding the study's test hypotheses, results were inconclusive. Only H1 (incentivized contract types correlate with higher vendor performance scores) and H3 (competed contracts correlate with higher vendor performance scores) were thoroughly evaluated, and proved to be inconclusive via initial standard regression technique. Due to datasets being too small for substantive use in big data network evaluation, or, because of time limitations preventing necessary dataset concatenation, H2 (shorter duration contracts correlate with higher vendor performance scores) and H4 (contract clauses have impact on vendor performance score) could not be evaluated.