Distributed agile development in multi-national projects brings with it a number of issues, including higher risk and decreased predictability. To solve these issues, this work introduces the Scrum-tree-k-nearest neighbor's algorithm, which combines Scrum principles with the k-nearest neighbor's method to dynamically assign resources in agile development projects. Scrum is a system for managing complex goods and solutions that is iterative and progressive. The k-nearest neighbour method is a machine learning approach that distributes resources dynamically depending on task proximity. The Scrum-tree-k-nearest neighbour method delivers a dynamic, flexible, and adaptable approach to project management by merging these two approaches. The study showcases the newly created algorithm and analyses 15 agile projects to demonstrate its usefulness. The findings show that the Enhanced Scrum-tree-k-nearest neighbour technique is more efficient and effective than standard software development life cycle approaches, and it aids in mitigating the hazards of distributed agile development. Finally, the Enhanced Scrum-tree-k-nearest neighbor's method provides a viable technique for improving the results of future agile development initiatives. The suggested method allows for dynamic resource allocation, which reduces risk and enhances predictability, boosting the overall efficiency and effectiveness of agile development projects.

Scrum, Agile Development, KNN, Resource Allocation, Sprints.