Software defect prediction using Bayesian networks
Empirical Software Engineering
| Hi-index | 0.00 |
Importance of software maintainability is increasing leading to development of new sophisticated techniques. This paper presentes the applications of support vector machine and unsupervised learning in software maintainability prediction using object-oriented metrics. In this paper, the software maintainability predictor is performed. The dependent variable was maintenance effort. The independent variable were five OO metrics decided clustering technique. The results showed that the Mean Absolute Relative Error (MARE) was 0.218 of the predictor. Therefore, we found that SVM and clustering technique were useful in constructing software maintainability predictor. Novel predictor can be used in the similar software developed in the same environment.