A Machine Learning Approach to Predicting Federal STEM Workforce Attrition

The federal government of the United States continuously needs employees with science, technology, engineering, and mathematics (STEM) skills to fill critical positions. Many federal agencies are failing to recruit and hire employees in sufficient numbers to accomplish their mission objectives. The compounding effect of inaccurate turnover forecasts has exacerbated the STEM workforce shortage in the federal government. The inability of human resources professionals to accurately predict STEM attrition has hampered agencies’ abilities to secure proper budgetary resources and successfully incorporate staffing predictions into annual hiring plans.This praxis tests the leading supervised machine learning classification models’ capacities to predict STEM workforce attrition. The data for this project were acquired from the U.S. Department of State and encompass 22 attributes of 5,442 employees who served in a STEM position between May 2011 and April 2021. This praxis also used the Department of State’s employee engagement and global satisfaction indices from the Federal Employee Viewpoint Survey, which is administered annually. The findings indicate that machine learning models can accurately predict attrition. The extra trees model was the most accurate (95.13%) in correctly predicting the majority class (employees who stayed) and the minority class (those who left) using the training and testing datasets.

Author

• Pasquarella, Albert

Language

• en

Date created

• 2023

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• Jones, Michael C

Committee Member(s)

• Blackford, Joseph P
• Etemadi, Amir

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/s7526d181