An Open-Domain Chinese Question Answering System Based on Dependency Parsing

摘要：

This paper presents an open-domain Chinese Questionanswering system mongoose, which aims to answer all kinds of factoid questions regardless of the theme and time. It first introduces our strategy for question classification using templates and HowNet, and then various techniques are discussed in the processing of the query sentence, including word segmentation, keyword extraction and NER (named entity recognition). An approach of combining the result of dependency parsing and keywords density is proposed in reranking of the sentences retrieved from a search engine. At last, we used four kinds of measurement to weighting the candidate answers and select out the top 10. Result of experiment conducted shows a good performance of our system.

关键词： question answering dependency parsing HowNet. sentence reranking

作者: Ningfeng Liu Xiaodong Shi

作者单位: Cognitive Science Department, Xiamen University, Xiamen, China, 361005 Fujian Key Laboratory of the Brain-like Intelligent Systems (Xiamen University), Xiamen, China, 361005

会议类型: 国际会议

会议名称: 2010 International Conference on Measurement and Control Engineering(2010年IEEE测量与控制工程国际会议 ICMCE2010)

会议地点: 成都

会议语种:英文

页码: 374-377

在线出版日期: 2010-11-16（万方平台首次上网日期，不代表论文的发表时间）