Software Defect Estimation using Support Vector Regression

By using a kernel function, data that are not easily separable in the original space can be predicted data. This paper evaluates the performance of support vector regression for software defect estimation. In addition, this method is compared with the two statistical methods: kernel and linear regression. The performance of the methods is assessed by the mean magnitude of relative errors (MMRE). Experiments were carried out using a project data set from NASA and the results show that support vector regression methods gives better performance than regression statistical methods in this task.