Enhancing Software Effort Estimation in the Analogy-Based Approach Through the Combination of Regression Methods

The success of software projects is closely linked to accurate effort estimation, driving continuous efforts by researchers to refine estimation methods. Among various techniques, the analogy-based approach has emerged as a widely-used method for software effort estimation. However, there is still a need to improve its accuracy and reliability. This study aims to enhance software effort estimation in analogy-based methods by introducing a hybrid approach that combines multiple regression methods with feature weighting. The proposed approach evaluates various regression models, integrating them with analogy-based estimation using a weighted combination of project features. The objective is to improve the precision of effort estimation by optimizing similarity functions and project attribute weights. Experimental results demonstrate that the hybrid model significantly outperforms traditional analogy-based methods, achieving superior accuracy across various software project datasets. The findings highlight the potential of this approach to offer a more dependable foundation for software effort estimation, contributing to the success of software projects.