声明
Abstract
摘要
CONTENTS
List of Figures and Tables
Table of Major Symbols and Units
1 Introduction
1.1 Thesis Statement
1.2 Contribution
1.3 Organization
2 Background and State of the Art
2.1 Mining Software Repositories
2.2 A Summarization Process:Common Steps
2.2.1 Corpus Creation
2.2.2 Normalization
2.2.3 Experimentation and Summary
2.2.4 Evaluation
2.3 Automatic Summarization
2.4 State of the Art
2.4.1 IR based Studies
2.4.2 NLP based Studies
2.4.3 Studies using Stereotype Identification
2.4.4 Program Analysis based Studies
2.4.5 Machine Learning based Studies
2.5 Applications of Software Artifact Summarization
3 Summarizing Bug ReportseoortS
3.1 Introduction
3.1.1 Hypothesis
3.1.2 Solution
3.2 Motivation
3.3 Corpus Collection
3.3.1 The MBRC Corpus
3.3.2 The OSCAR Corpus
3.3.3 The Annotation Process
3.4 Bug Report Summarizer-PRST
3.4.1 Sentence Splitter
3.4.2 Ranking Module
3.4.3 Regression Module
3.4.4 Prediction Module
3.4.5 Ranking Merger
3.4.6 PRST Algorithm
3.4.7 Features
3.4.8 Example Summary
3.5 Analytical Evaluation
3.5.1 Performance Evaluation
3.5.2 Threats to the Validity
3.6 Summary
4 Code Fragment Summarization
4.1 Introduction
4.2 Approach
4.2.1 Corpus Creation
4.2.2 Corpus Annotation
4.2.3 Feature Extraction through Crowdsourcing
4.3 Code Fragment Summarizer
4.4 Algorithm for CFS
4.5 Evaluation
4.5.1 Statistical Evaluation
4.5.2 Feature Selection Analysis
4.6 Summary
5 Discussion and Future Work
5.1 Discussion
5.1.1 Producing Bug Report Summaries
5.1.2 Producing Source to Source Summaries
5.1.3 Type of Summary Generated
5.1.4 Bug Reports and Mailing Lists
5.1.5 Summary Perspective
5.2 Future Directions
5.2.1 Crowdsourcing
5.2.2 Heterogeneous Artifacts
5.2.3 Industry and Academia Collaboration
Conclusion
Abstract of Innovation Points
References
Published Academic Papers during PhD Period
Acknowledgement
About the Author