An Optimized SVM Model for Accurate Duplicate Bug Report Retrieval

Bug reporting facility is an integral part of software support system and a critical aspect of beta testing. Managing the incoming bug reports for large software projects is a challenging task. Software testers or end users submit bug reports, as and when potential defects are identified in the system. Sometimes two or more bug reports corresponds to the same defect. The analysis of reports to check whether the incoming report is for the same bug for which a master report already exists in the database is one of the most important task in managing the bug reports. This analysis needs to be done very carefully so that a report corresponding to a bug which is already reported is discarded while at the same time, a textually similar report, but corresponding to a different bug is to be attended properly. To address this problem with duplicate bug reports, a person called a triager needs to manually label these bug reports as duplicates, and link them to their "master" reports for subsequent maintenance work. However, for large software systems, this process could be highly time consuming. To address this issue, recently, several techniques have be proposed using various similarity based metrics to detect candidate duplicate bug reports for manual verification. Automating triaging has been proved challenging as two reports of the same bug could be written in various ways. There is still much room for improvement in terms of accuracy of duplicate detection process. In this dissertation, an optimized SVM model is proposed to detect duplicate bug reports more accurately. The proposed techniques are validated on three large software bug repositories from Firefox, Mysql, and Apache. Analytical methods are presented over samples and results are provided which are in excellent agreement with the simulation results.