ECLogger: Cross-Project Catch-Block Logging Prediction Using Ensemble of Classifiers

Background: Software developers insert log statements in the source code to record program execution information. However, optimizing the number of log statements in the source code is challenging. Machine learning based within-project logging prediction tools, proposed in previous studies, may not be suitable for new or small software projects. As these software projects do not have sufficient prior training data to construct the prediction model. For such software projects, we can use cross-project logging prediction. Aim: The aim of the study presented is paper is to investigate cross-project logging prediction methods and techniques. Method: We propose ECLogger, which is a novel, ensemble-based, cross-project, catch-block logging prediction model. We use 9 base classifiers and combine them using 3 ensemble techniques to improve the cross-project logging prediction result. We evaluate the performance of ECLogger on three open-source Java projects: Tomcat, CloudStack and Hadoop. Results: ECLoggerBagging, ECLoggerAverageV ote, and ECLoggerMajorityV ote show a considerable improvement in the average LF in 3, 5, and 4 source→target project pairs, respectively, compared to the baseline classifiers. ECLoggerAverageV ote performs the best and shows improvements of 3.12% (average LF) and 6.08% (average ACC) compared to the baseline classifier. Conclusion: The classifier based on ensemble techniques such as bagging, average vote and majority vote outperforms the baseline classifier. Overall, the ECLoggerAverageV ote model performs the best.The results show that the CloudStack project is more generalizable than the other projects.