Prediction of Story Point Estimation with Transformer-Based Architecture and Machine Learning Models

Pandey, Purnima (2024) Prediction of Story Point Estimation with Transformer-Based Architecture and Machine Learning Models. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (3MB) | Preview
Preview	PDF (Configuration Manual) Download (3MB) | Preview

Abstract

Estimation of story points is highly valuable in Agile methodology in that it organizes the way resources are utilized, when they are to be used and the time that is available for the project to be completed. While the classical estimation techniques rely primarily on the expert’s experience and judgment, it can prove to be rather infeasible and lead to total inconsistency in the Agile process that, naturally, will impact the success of the Agile project. Nevertheless, the current state in machine learning approaches used in Agile development is still inadequate as per the complexity of models that can be applied. This study seeks to make that literature review by employing transformer-based architectures more specifically GPT-2SP model to predict story points in Agile projects. The dataset of 23313 issues in 16 open source projects was used to test several machine learning algorithms of story point estimation including SVM, KNN, RF, GBM, and LR. Therefore it was ascertained that in as much as the evaluation results on KNN showed that KNN has the capacity to give the best accuracy possible from 87% to 89%, especially when grouped based on the importance of tasks. The GPT-2SP model also uncovered lots of potential by reducing the bias introduced by the typical method which relies heavily on analysts’ estimate to specifically identify the numerical values and the results appear to be very close and less scattering than the conventional one. Such outcomes call for the possible application of Machine Learning models in the Agile management of projects, in view of the enhanced predictive accuracy, effective decision-making and enhanced efficiency in the right resource utilisation. As a result, Agile teams manage to manage the project scope, reduce the subjectivity of estimation in overall team productivity and efficiently increase the scale of the project.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Rustam, Furqan UNSPECIFIED
Uncontrolled Keywords:	Story Point Estimation; Transformer-Based Architecture; Agile Project Management; GPT-2SP Model; Effort Estimation
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management > Project management
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	25 Aug 2025 09:03
Last Modified:	25 Aug 2025 09:03
URI:	https://norma.ncirl.ie/id/eprint/8604

Actions (login required)

View Item