Automated Bug Triage using Advanced Data Reduction Techniques
Author(s):
Irin Ani John , Amal Jyothi College of Engineering, Kanjirappally, Kottayam, India; Tintu Alphonsa Thomas, Amal Jyothi College of Engineering, Kanjirappally, Kottayam, IndiaKeywords:
Data mining, bug triage, classification, data reduction, instance selection, feature selectionAbstract:
Bug triage is an important step in the process of bug fixing. The goal of bug triage is to correctly assign a developer to a newly reported bug in the system. To perform the automated bug triage, text classification techniques are applied. This will helps to reduce the time cost in manual work. To reduce the scale and improve the quality of bug data, the proposed system addresses the data reduction techniques, instance selection and feature selection for bug triage. The instance selection technique used here is to identify the relevant bugs that can match the newly reported bug. The feature selection technique is used to select the relevant data from each bug in the training set. A predictive model is proposed to identify the order in which the data reduction techniques are applied for each newly reported bug. This step will improve the performance of the classification process. An experimental study using Eclipse and Firefox bug data is undergone in which the proposed system shows an accuracy of 73%.
Other Details:
Manuscript Id | : | J4RV2I6009 |
Published in | : | Volume : 2, Issue : 6 |
Publication Date | : | 01/09/2016 |
Page(s) | : | 24-27 |