Portal de Periódicos da CAPES

Neural Network Based Bug Priority Prediction Model Using Text Classification Techniques

2017; International Journal of Advanced Research in Computer Science; Volume: 8; Issue: 5 Linguagem: Inglês

10.26483/ijarcs.v8i5.3559

0976-5697

Pooja Choudhary,

Software Engineering Techniques and Practices

As open source software systems are becoming larger and more complicated, the task of bug detection and fixing to improve the performance of the software is also getting more complex, time consuming and inefficient. To improve the quality and efficiency of the software, developers allow users to report bugs that are found by them using bug tracking system such as Bugzilla. In Bugzilla users specify the details of the bug, such as the description, the component, the version, the product and the severity. Depending on this information the developers assign the priority levels to the reported bugs. The task of prioritizing the bug reports is manual, therefore it is time-consuming and inconsistent. In this dissertation, Neural Network technique is used for developing prediction models for five different versions (2.0, 2.1, 3.0, 3.1, and 3.2) of Eclipse that will assign the priority levels based on the information provided in each bug report. The features that potentially affect the priority of a bug are temporal, textual, author-related, severity, product and component. The collected dataset is used to train and test the classification algorithms. ROC and F-measure are used to interpret the results.