A GA-based feature selection and parameters optimization for support vector regression applied to software effort estimation

The precision of the estimation of the effort of software projects is very important for the competitiveness of software companies. Machine learning methods have recently been applied for this task, included methods based on support vector regression (SVR). This paper proposes and investigates the use of a genetic algorithm approach for simultaneously (1) select an optimal feature subset and (2) optimize SVR parameters, aiming to improve the precision of the software effort estimates. We report on experiments carried out using two datasets of software projects. In both datasets, the simulations have shown that the proposed GA-based approach was able to improve substantially the performance of SVR and outperform some recent results reported in the literature.