Empirical Assessment of Machine Learning Techniques for Software Requirements Risk Prediction

Software risk prediction is the most sensitive and crucial activity of Development Life Cycle (SDLC). It may lead to success or failure a project. The should be predicted earlier make software project successful. A model proposed for requirement risks using dataset machine learning techniques. In addition, comparison made between multiple classifiers that are K-Nearest Neighbour (KNN), Average One Dependency Estimator (A1DE), Naïve Bayes (NB), Composite Hypercube on Iterated Random Projection (CHIRP), Decision Table (DT), Table/Naïve Hybrid Classifier (DTNB), Credal Trees (CDT), Cost-Sensitive Forest (CS-Forest), J48 Tree (J48), (RF) achieve best suited technique according nature dataset. These techniques evaluated various evaluation metrics including CCI (correctly Classified Instances), Mean Absolute Error (MAE), Root Square (RMSE), Relative (RAE), Squared (RRSE), precision, recall, F-measure, Matthew’s Correlation Coefficient (MCC), Receiver Operating Characteristic Area (ROC area), Precision-Recall Curves area (PRC accuracy. inclusive outcome this study shows in terms reducing error rates, CDT outperforms other achieving 0.013 MAE, 0.089 RMSE, 4.498% RAE, 23.741% RRSE. However, increasing accuracy, DT, DTNB, better results.