NEW SOFTWARE DEFECT PREDICTION METHOD BASED ON PCA AND OPTIMIZED LSTM
It is natural for the developed software to contain some defects and errors. The important issue is to detect these errors. Tests play an important role in detecting errors, but this is not always sufficient. Therefore, effective methods are needed to detect software errors and defects. The detecting error before publishing and serving the software the is main aim of the companies and the software engineers but this issue remain difficult and challenging issue. In this study, new method based KNN, and parameter optimization techniques presented in the field of software defect prediction. The proposed method applied ensemble learning with KNN based Grid search to present new model for software defect detection. Furthermore, several classifiers are applied to obtain best detection rate. The obtained results are compared and discussed to select best model. The presented method obtained suitable results when compared with traditional techniques. The three datasets that are used in this study are CM1, JM1, and PC1 which proposed method presented various results with these datasets vary between 75% to 94%.
---