In Agile software development, effective task prioritization is essential for ensuring timely delivery and efficient resource management throughout the software development life cycle. Traditional prioritization methods often depend on subjective decision-making, which can lead to inefficiencies. This paper discusses the challenges associated with conventional prioritization methods. It presents an AI-driven approach to automating the process, addressing new real-world challenges such as reducing subjectivity and adapting to changing requirements. The proposed methodology includes data collection, model selection, training the model, and tuning model parameters in machine learning models. It incorporates textual descriptions where feature extraction is essential, utilizing pre-trained Transformer models alongside structured features. These models will be evaluated using performance metrics, and inferences will be drawn from the results obtained. The results to be obtained are Task ranking, which has continuous values. It is a Regression model, where Linear Regression is considered a Base model compared with a tree-based model such as Random Forest and the XGBoost model. Our results demonstrate which XGBoost model has the strongest accuracy of 97%.In contrast, the Random Forest has 91%, and the Base model has 87% test accuracy. The Task is to be performed with high priority with the help of prioritization techniques, considering the specific factors to be fulfilled for the successful completion of the project across the organization. Explainable AI techniques, such as SHAP and LIME, will be implemented to enhance task ranking, transparency, and explainability. These techniques help interpret model predictions and highlight the contribution of each feature to the final rankings. The proposed approach improves task prioritization automation by reducing manual effort, enhancing decision accuracy and scalability, and ultimately contributing to the efficiency of Agile software development. This work contributes to the intersection of Agile software development and AI, paving the way for intelligent Agile project management and opening new avenues for AI-driven software engineering.
Please read the paper in the 'paper' folder for the details. Also, the paper has not been sent to a peer-reviewed venue.