Write a Blog >>
MSR 2022
Mon 23 - Tue 24 May 2022
co-located with ICSE 2022

Background: Selecting a suitable feature reduction technique, when building a defect prediction model, can be challenging. Different techniques can result in the selection of different independent variables which have an impact on the overall performance of the prediction model. To help in the selection, previous studies have assessed the impact of each feature reduction technique using different datasets. However, there are many reduction techniques, and therefore some of the well-known techniques have not been assessed by those studies. Aim: The goal of the study is to select a high-accuracy reduction technique from several candidates without preliminary assessments. Method: We utilized bandit algorithm (BA) to help with the selection of best features reduction technique for a list of candidates. To select the best feature reduction technique, BA evaluates the prediction accuracy of the candidates, comparing testing results of different modules with their prediction results. By substituting the reduction technique for the prediction method, BA can then be used to select the best reduction technique. Results: In the experiment, we evaluated the performance of BA to select suitable reduction technique. We performed cross version defect prediction using 14 datasets. As feature reduction techniques, we used two assessed and two non-assessed techniques. Using BA, the prediction accuracy was higher or equivalent than existing approaches on average, compared with techniques selected based on an assessment. Conclusions: BA can have larger impact on improving prediction models by helping not only on selecting suitable models, but also in selecting suitable feature reduction techniques.

Fri 20 May

Displayed time zone: Eastern Time (US & Canada) change

04:00 - 04:50
Session 14: Software Quality Technical Papers / Industry Track / Data and Tool Showcase Track at MSR Main room - even hours
Chair(s): Chakkrit Tantithamthavorn Monash University, Simone Scalabrino University of Molise
04:00
4m
Short-paper
Evaluating the effectiveness of local explanation methods on source code-based defect prediction models
Technical Papers
Yuxiang Gao Jiangsu Normal University, Yi Zhu Jiangsu Normal University, Qiao YU Jiangsu Normal University
Pre-print
04:04
7m
Talk
Problems and Solutions in Applying Continuous Integration and Delivery to 20 Open-Source Cyber-Physical Systems
Technical Papers
Fiorella Zampetti University of Sannio, Italy, Vittoria Nardone University of Sannio, Massimiliano Di Penta University of Sannio, Italy
04:11
7m
Talk
To Type or Not to Type? A Systematic Comparison of the Software Quality of JavaScript and TypeScript Applications on GitHub
Technical Papers
Justus Bogner University of Stuttgart, Institute of Software Engineering, Empirical Software Engineering Group, Manuel Merkel University of Stuttgart
Pre-print
04:18
7m
Talk
Using Bandit Algorithms for Selecting Feature Reduction Techniques in Software Defect Prediction
Technical Papers
Masateru Tsunoda Kindai University, Akito Monden Okayama University, Koji Toda Fukuoka Institute of Technology, Amjed Tahir Massey University, Kwabena Ebo Bennin Wageningen University and Research, Keitaro Nakasai National Institute of Technology, Kagoshima College, Masataka Nagura Nanzan University, Kenichi Matsumoto Nara Institute of Science and Technology
Pre-print
04:25
4m
Talk
Constructing Dataset of Functionally Equivalent Java Methods Using Automated Test Generation Techniques
Data and Tool Showcase Track
Yoshiki Higo Osaka University, Shinsuke Matsumoto Osaka University, Shinji Kusumoto Osaka University, Kazuya Yasuda Hitachi, Ltd.
Media Attached
04:29
7m
Talk
Extracting corrective actions from code repositories
Industry Track
Yegor Bugayenko Huawei, Kirill Daniakin Innopolis University, Mirko Farina Innopolis University, Firas Jolha Innopolis University, Artem Kruglov Innopolis University, Witold Pedrycz University of Alberta, Giancarlo Succi Innopolis University
04:36
14m
Live Q&A
Discussions and Q&A
Technical Papers


Information for Participants