• Sinduja.B
• P.M. Kamatchi

Data mining can be used to explore interesting patterns in software bug repositories. Managing large quantity of new bugs submitted by testers/reporters on daily basis in large organizations is a complex task. This increases the defect triage process, where the bugs are analyzed and prioritized within specified period to assign appropriate developers to fix the defects based on the severity level of the bugs. In this paper, a prediction model is proposed to predict complexity and severity level of bugs. Complexity and severity level of a bug helps the development and testing team in an organization to plan future software releases. For all the bugs stored in bug repository, the time required to fix the bug i.e., fix duration/time is calculated, and complexity clusters (simple, medium, and complex) are created for the computed fix duration. The bugs are then mapped to the complexity cluster based on the estimated fix duration. In this study, we have extracted bug report dataset from Bugzilla bug tracking system. We compare four algorithms, namely, Naïve Bayes, Decision Tree, Random Forest, and Convolutional Neural Network for Classification and parameters measured in terms of accuracy and F-score is evaluated.

Bug Complexity, Clustering, Prediction, Accuracy

"Predicting Software Defect Complexity and Accuracy using Bug Estimation and Clustering", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.9, Issue 3, page no.e542-e547, March-2022, Available :http://www.jetir.org/papers/JETIR2203470.pdf

"Predicting Software Defect Complexity and Accuracy using Bug Estimation and Clustering", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.9, Issue 3, page no. ppe542-e547, March-2022, Available at : http://www.jetir.org/papers/JETIR2203470.pdf