Header menu link for other important links
Application of SMOTE and LSSVM with Various Kernels for Predicting Refactoring at Method Level
Kumar L, , Krishna A.
Published in Springer International Publishing
Volume: 11305 LNCS
Pages: 150 - 161
Improving maintainability by refactoring is essentially being considered as one of the important aspect of software development. For large and complex systems, identification of code segments, which require re-factorization is a compelling task for software developers. Development of recommendation systems for suggesting methods, which require refactoring are achieved using this research work. Materials and Methods: Literature works considered source code metrics for object-oriented software systems in order to measure the complexity of a software. In order to predict the need of refactoring, the proposed system computes twenty-five different source code metrics at the method level and utilize them as features in a machine learning framework. An open source dataset consisting of five different software systems is being considered for conducting a series of experiments in order to assess the performance of proposed approach. LSSVM with SMOTE data imbalance technique are being utilized in order to overcome the class imbalance problem. Conclusion: Analysis of the results reveals that LS-SVM with RBF kernel using SMOTE results in the best performance. © 2018, Springer Nature Switzerland AG.
About the journal
JournalData powered by TypesetNeural Information Processing Lecture Notes in Computer Science
PublisherData powered by TypesetSpringer International Publishing
Open Access0