Effort Estimation for Redmine Tickets Using Machine Learning

This paper presents the application of machine learning for estimating effort in software projects, specifically focusing on Redmine tickets. It includes data collection, preprocessing, and model identification, resulting in a stable predictive model achieving a PRED 90% prediction accuracy using real-world data. Incorporating solo and ensemble techniques, the paper compares their performances and provides methodologies for precise effort estimation, supporting project managers in early-stage planning. Project managers can benefit by adopting new estimation methods, fostering improved productivity and client relationships. We also address challenges such as evolving programmer’s productivity, system complexity, and model accuracy fluctuations over time, emphasizing the necessity for continuous training and optimization with updated data to ensure relevance and precision.