Software Fault Prediction Using Machine Learning Models

Abstract

In recent years, computers have great role to the society for their reliability which becoms a key essential in day to day life. The role of software and its captious function in computer system for some certain software has appeared as important achievement for certain infrastructure. Exploitation of system perspective which recognise the importance of software that characterized the current state of fault identification research work as it contributes to the reliability of computer systems. In general, different classification algorithms like K-Nearest Neighbors (KNN), Decision Tree (DT), Naive Bayes (NB), Radial Basis Function Support Vector Machine (RBF-SVM), (L-SVM), Polynomial Support Vector Machine (P-SVM), Adaboost, and Random Forest (RF) have been considered to determine classification performance to evaluate the accuracy of classification with ten number of fault-tolerance datasets. In most of the cases, it is noticed that the nature of data have great impact in the performance of the classification algorithm. The evaluation of several performance measures of all the above ML classification algorithms have been analyzed for ten number of fault-tolerance datasets. It is also observed that the classifier Adaboost gives better result as compared to rest of the classification algorithms.