Machine Learning Model for the U.S. Customs and Border Protection – Cargo Systems Program Directorate’s Sprint Effort Capacity Estimation
Open AccessAgile methodology has been widely adopted by both commercial and government software development projects since 2001. Agile methodology promotes product delivery by executing multiple small iterations that are also known as sprints. Each sprint is a small software development project with own planning, development, testing, demonstration, with possible deployment for production. Agile software development projects commonly use Yesterday’s Weather Model to estimate sprint effort capacity. However, the accuracy of Yesterday’s Weather Model is unreliable. Over 60% of Agile software development projects experience schedule delays, cost overruns, or cancellations and inaccurate effort estimation is one of the leading causes of these issues. As such, Agile software development projects may benefit from a sprint effort capacity estimation model with improved accuracy. In this research, we compute the error rate of Yesterday’s Weather Model using a large-scale real data from the U.S. Customs and Border Protection – Cargo Systems Program Directorate and identify a list of essential predictors that can be used to estimate sprint effort capacity. Using machine learning algorithms, we develop, test, and validate a sprint effort capacity estimation model on the same historical data. The model demonstrated better performance when compared to other models including Yesterday’s Weather Model.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.