Title: Automated Software Bug Localization
Abstract: A software bug is a coding mistake that may cause an abnormal behavior of the software product. When a user observes an abnormal behavior of the product, he will report it in a document called bug report. When a new bug report is received, developers usually need to reproduce the bug and perform code reviews to find the cause. This manual process is tedious and time consuming. To help software developers find software bugs more efficiently, we introduce a ranking model to rank all the source code files for a given bug report automatically. The higher position in the ranked list indicates the larger probability that the source file contains the bug. The ranking model is a weighted combination of various features. Every feature is a type of information that measures the relationship between a source code file and the bug report. The weight parameters of different features are learned automatically based on previously solved reports. From intensive evaluations we found that this ranking model outperformed the baseline approaches. Furthermore, we can always create and add new features to achieve better ranking performance.
Bio: Xin Ye received the PhD degree in Computer Science from Ohio University. He is currently an Assistant Professor in California State University San Marcos. His main research interests are in software engineering and machine learning, with a recent focus on mining software repository to support software maintenance and applying software engineering principles in game design.