作者: S. J. Young , H. -H. Shih
DOI: 10.1007/3-540-58473-0_156
关键词:
摘要: This paper presents a system for computer assisted grammar construction (CAGC). The CAGC is designed to generate broad-coverage grammars large natural language corpora by utilizing both an extended inside-outside algorithm and automatic phrase bracketing (AUTO) system, which provide the with constituent information during learning. demonstrates capability of deal realistic problems usefulness AUTO in based re-estimation. Performance results including analysis degree coverage precision are presented constructed Wall Street Journal (WSJ) corpus.