應用於中文文件擷取的語言查詢模式設計與製作

Abstract

在行政院大力推動國家資訊基礎建設下, 中文資訊網路日益普及於各行各 業。因此, 中文資訊查詢系統的研發也就日益重要。有鑑於此, 本論文將 探討如何設計一個自然且便利的語言查詢模組, 適用於中文電子文件擷取 系統。我們所設計的模組將利用模糊理論來建立語意變數(linguistic variable) 的語法結構, 藉以描繪關鍵詞的重要程度。另外我們也設計了 巨集運算元(aggregation operator),來簡化複雜的佈林模式查詢法語意 中的不明確性。最後提出改良式組合查詢法, 實驗結果驗證所提的方法比 傳統模糊理論的處理提供較佳的查詢滿意度。 Chinese information network has been popular in the society of allareas under the promotion of NII project. Hence, the research in the Chinese information retrieval system becomes more important thanever. In this thesis, we will investigate the design of a user-friendlylinguistic model particulary for Chinese document retrieval system.Such model is constructed based on fuzzy set theory and constructedwith linguistic variables and their syntactic structure to describlethe degree of important for each keyword. In addition, we also designaggregation operators that can reduce the ambiguity in the complexBoolean model. Finally, we propose the improved combination query method.The experimental results show that our method yield better degree ofrequest satisfaction than the method based on traditional fuzzy theory.