Mining data from the army reserve for analysis of attrition factors

Abstract

The goal of this thesis was to analyze the impact of increased utilization and deployments of Troop Program Unit soldiers since 9/11, countered against the effects of demographics and of the programs and actions meant to control attrition. This study conducted a process of data collection, data manipulation, and data-mining algorithms executed against the entire enlisted TPU population and focused toward attrition behavior. Significant factors in determining attrition behavior included time in service, increased bonus levels and the Delayed Entry Program. Mobilizations, in and of themselves, appear to have little impact. The models we built showed significant potential for predicting behavior. We believe that this process should be continued and expanded to a tool to aid in and affect attrition.