Classification of Bug Report Using Naïve Bayes Classifier with Gain Ratio
Abstract
Bug report is a report which contains the information about the defects in the system or in the software. Generally, bug report contains the issues written by the wide variety of reporters, with different levels of training and knowledge about the system being discussed. Bug tracking systems are made to manage bug reports, which are collected from various sources. These bug reports are needed to be labeled as security bug reports or non security bug reports, since security bug reports (SBRs) contain more risk than non-security bug reports (NSBRs). In this paper we are using Naive Bayes classifier to classify the bug reports. With naive bayes classifier, feature subset selection method such as Gain Ratio is applied to rank the attributes of the dataset. Gain Ratio is utilized as an iterative process where we select smaller sets of features in incremental manner. Result prove that the classification accuracy is high for attributes having high gain ratio and low for attributes having low gain ratio.
Authors
Smita Mishra , Somesh Kumar