（北京航空航天大學 軟件學院北京市 100191）

摘要：隨著互聯網和信息檢索技術的高速發展,近年來問答系統的應用日趨成熟，語義搜索引擎、語音助手等新產品大量涌現。配合語音識別技術，問答系統作為一種對現有搜索引擎的優化也能滿足用戶更豐富的信息檢索需求。以Google為代表的第二代搜索引擎已經實現了對關鍵詞相關的海量信息的快速檢索，而這也帶來了用戶在海量搜索結果面前的“信息迷失”。希望通過自然語言描述,表達自己的查詢需求,希望搜索服務系統能夠理解用戶意圖,返回恰當的結果，因此更加符合用戶檢索需求的技術應運而生，語義搜索引擎、問答系統的研究成為當前自然語言處理領域中最有活力的方向之一。廣域問答系統具有非常廣泛的應用前景，例如網絡答疑、公司客服等方面。本論文以研制廣域的問答系統為目標，根據國內外問答系統的發展現狀和所取得的成果,對中文問答系統及自然語言處理的相關的關鍵技術進行了較為深入的分析與探討。經過技術選型，系統設計和實現，提供了一種良好的用戶信息檢索解決方案。
關鍵詞：問答系統；問句處理；答案抽��；語義相似度。
Design and Implementation of Chinese Question-Answering System Based on Sematic Comprehension
SicongCheng
（School of Software, Beijing University of Aeronautics and Astronautics, Beijing 100191）
vheaven@163.com
Abstract：With the deeply development of internet and the rapid development of information retrieval technology in the world, the application of QA system matures, semantic search engines, voice assistant, and other new products are growing in large numbers. With voice recognition technology, answering system as an existing search engine optimization can meet user demand for richer information retrieval. Represented by Google's second-generation searching engine has achieved a mass of information on Keywords rapid retrieval, but it also brings the user in front of massive search results "message Lost". Hope that, through natural language and express their query requirements, they want the search service system to understand user intent, returns the appropriate result, and therefore more in line with user demand for technology emerges, semantic searching engines, QA system has become the NLP in the field one of the most dynamic direction. QA system has a very wide range of potential applications such as network answering, customer service and other aspects. In this thesis, the wide-area targeted QA system, according to the current development of domestic and abroad QA system and the results obtained on the Chinese QA system and analyses key techniques of Chinese QA and question processing deeply. After technology selection, system design and implementation,a good user information retrieval solution is provided.
Key words：QA system, Semantic similarity, Question processing, Answer extraction.

參考文獻
（1）“Lemur toolkit for language modeling and information retrieval,” http://www.lemurproject.org/.
（2）馬金山，基于統計方法的漢語依存句法分析研究.
（3）段建勇，現代漢語詞性細分類標注研究，山西大學碩士學位論文，山西，山西大學，2004
（4）張會鵬，中文詞法分析技術的研究與實現，哈爾濱工業大學碩士學位論文，黑龍江，哈爾濱工業大學，2006
（5）文勖，張宇，劉挺，類別主特征結合句法特征的中文問題層次分類，中國中文信息學會信息檢索與內容安全專業委員會，第二屆全國信息檢索與內容安全學術會議（NCIRCS-2005）論文集，中國北京，2005
（6）劉懷軍，車萬翔，劉挺，中文語義角色標注的特征工程，中文信息學報，2007，21，1，79-84
（7）V. Punyakanok，D. Roth，W. Yih，The Necessity of Syntactic Parsing for Semantic Role Labeling，In:Proceedings of CoNLL204，2004
（8）Zhang，D.Lee，W.， Web based pattern mining and matching approach to question answering，Proceedings of the 11th Text Retrieval Conference ，2002
（9）SEMANTIC EXTENSIONS OF THE EPHYRA QA SYSTEM FOR TREC 2007
（10）J. RuppeNpofer, M. Ellsworth, M.R.L. Petruck, C.R. Johnson, and J. Scheffczyk, “FrameNet II: Extended theory and practice,” http://framenet.icsi.berkeley.edu/