Defects Prediction Using Oo Metrics And Machine Learning To Improve Software Quality

In software development life cycle,software testing is the main stage,which can minimize the defects of software.A domain that has receiving much attention of software researchers since past couple of years is software defects prediction(SDP).Its aim to minimize the cost,time and improve the efficiency of software.Software defect prediction has become more and more important in software reliability in the last couple of years.However,on software testing,we are wasting much time,resources,and money.Software defect prediction can help to improve the efficiency of software testing and guide direct resource allocation.Researchers and software engineers have proposed many defect prediction models.However,until now,no reasonable outcome to prove which metrics set or classifier has the best result so far.To enhance defect model accuracy,the researchers applied various machine-learning classifiers to predict defects.The main aim of this research is to show a comparative analysis of software defects prediction based on support vector machine SVM and extreme learning machine ELM.In this domain defect,prediction models usually created using three different prediction techniques based on test data and training data.i.e.cross-validation prediction,cross-version prediction and cross-project prediction.In our study,we used cross version prediction approach,data from old version of a software used as training data to develop the prediction model and the model evaluated from same project of current version.We collected three different versions of Eclipse version control system,from the development of Eclipse program repository.The very first stage to develop a prediction model is to create instances from software records,such as version control systems in short VCS,bug-tracking systems,email records,etc.An instance has many metrics or attributes extracted from the repositories of the software and is marked with full of defects / clean or number of defects.After creating instances with metrics and labels,which are popular in machine learning,we can introduce preprocessing steps.After pre-processing,we choose different object oriented metrics and algorithm to build our model,aiming to predict software defects in different versions.We split the data into file and packaged based.For training purpose of our model,we used SVM and ELM.To validate the performance of prediction model using some popular used measurement scales such as accuracy,precision,recall,AUC(Area under ROC curve).By comparing,the file based and package based result of SVM and ELM.The results demonstrate that support vector machine is best fit for the cross-version defect prediction.