A fullstack ecommerce app built in react
Loading...
Files
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
NHCE
Abstract
Software problems cost software businesses more than 45 percent of their revenue. Bug triage is an unavoidable stage in the bug fixing process, and it seeks to accurately allocate a developer to a new problem. Text categorization algorithms are used to conduct automatic bug triage to reduce the time cost of manual labor. We address the topic of data reduction for bug triage in this work, i.e., how to minimize the scale and increase the quality of bug data. We integrate instance and feature selection to minimize data size on both the bug and word dimensions at the same time. We gather variables from prior bug data sets and develop a prediction model for a new bug data set to decide the sequence of applying instance selection and feature selection. We empirically analyze data reduction efficiency on 600,000 bug reports from two big open-source projects, Eclipse and Mozilla. The findings suggest that our data reduction may efficiently reduce the data size while improving bug triage accuracy. Our study proposes a method for utilizing data processing techniques to generate reduced and high-quality bug data in software development and maintenance.