A Comparative Analysis of Different Machine Learning Techniques Used in Software Effort Estimation

Software effort estimation is a vital component of project management, encompassing the prediction of time, cost, and resources necessary for software development. Accurate effort estimation plays a pivotal role in effective project planning, resource management, and budgeting within the realm of software engineering. However, software effort estimation is challenging due to project complexity, evolving requirements, and the unique nature of each development environment. In recent years, various machine learning techniques, including regression, fuzzy logic, support vector machines, decision trees, genetic algorithms, artificial neural networks, ensemble methods, clustering, K-nearest neighbors, logistic regression, random forests, and hybrid approaches—have been employed to improve estimation accuracy.This research delivers a comparative evaluation of these machine learning methods, considering performance metrics such as mean absolute error, accuracy, root mean square error, mean relative error, usability, and scalability. Special attention is given to ensemble methods, which often outperform individual models in accuracy. The findings offer guidance for researchers and practitioners in selecting optimal machine learning techniques for software effort estimation, ultimately enhancing tool development and modeling accuracy for sustainable software development.