【摘要】：機械產(chǎn)品設計未來的發(fā)展趨勢之一是智能化,這就需要具有知識處理能力的專家系統(tǒng)通過人工智能技術快速有效地收集并分析用戶的需求信息,將用戶需求轉化為產(chǎn)品概念設計要求,從而設計出滿足用戶需求的產(chǎn)品。 本文將人工智能語言信息處理領域的核心課題與技術——自然語言理解技術應用于機械產(chǎn)品設計中,討論并研究了基于篇章的名詞省略現(xiàn)象及其恢復策略。針對在省略中占82%的名詞主語省略、賓語省略兩種類型,筆者分別從自然語言理解的句法分析、語義理解和篇章分析三個層面進行研究,并結合對語料庫實證性研究的統(tǒng)計數(shù)據(jù),總結段落范圍內主語省略和賓語省略的一般類型的恢復框架和六種特殊類型的恢復策略,利用篇章理解的主題信息對恢復的正確性進行驗證,從而設計出名詞省略恢復的整體模型。 最后,把名詞省略恢復模塊整合到專家系統(tǒng)中,應用于現(xiàn)代機械產(chǎn)品設計中,并對用戶需求取得較好的理解結果,為后續(xù)產(chǎn)品設計的標準化和智能化提供支持。
[Abstract]:One of the developing trends of mechanical product design in the future is intellectualization, which requires expert system with knowledge processing ability to collect and analyze user demand information quickly and effectively through artificial intelligence technology. The user requirement is transformed into the product conceptual design requirement, and the product that meets the user's demand is designed. In this paper, the core subject and technology of artificial intelligence language information processing, natural language understanding technology, is applied to mechanical product design, and the phenomenon of noun ellipsis based on text and its recovery strategy are discussed and studied. In view of the two types of noun subject ellipsis and object ellipsis which account for 82% of the ellipsis, the author studies them from three aspects: syntactic analysis, semantic understanding and text analysis. Combined with the statistical data of corpus empirical study, the paper summarizes the general types of restoration framework of subject ellipsis and object ellipsis in paragraph scope and six special types of recovery strategies. The correctness of restoration is verified by the topic information of text comprehension, and a global model of noun ellipsis restoration is designed. Finally, the noun ellipsis recovery module is integrated into the expert system and applied to the modern mechanical product design, and the result of understanding the user's requirements is obtained, which provides the support for the standardization and intelligence of the subsequent product design.
【學位授予單位】：西安電子科技大學
【學位級別】：碩士
【學位授予年份】：2011
【分類號】：TP391.1;TH122

【引證文獻】

中國碩士學位論文全文數(shù)據(jù)庫 前1條

1 嚴羽;自然語言理解中并列名詞歧義消解及其在智能儀器設計領域的應用[D];西安電子科技大學;2011年

，

本文編號：2380136