PARSING JAVA METHOD NAMES FOR IMPROVED SOFTWARE ANALYSIS
Sana Ambreen Malik · 2011
Modern software engineering tools are driven by sophisticated automatic software analysis. Previous research indicates that the natural language (user-defined names) provides strong lexical clues about program behavior and structure and can be used to increase the effectiveness of various software engineering tools. Automatic analysis of the natural language usage in software requires accurate automatic parsing of the multi-word names (e.g., isPointInImage). This research focuses on developing the analysis techniques for an accurate parser for multi-word Java method names; this includes both a part-of-speech tagger and phrase chunker. These contributions form the foundation for natural language program analysis.