資訊檢索系統設計的目的是希望使用者能藉以擷取所需的資訊，同時過濾無用的 資訊。系統文件與查詢問句所具有的相同概念可經由詞彙顯現，如果檢索者可以 將資訊需求以正確且適當的詞彙表達，而系統也可以將之轉換為相同溉念的詞彙， 則檢索結果應能夠滿足使用者的資訊需求。然而，使用者很難建構出適當的查詢 問句，解決的辦法之一是擴展查詢問句，亦即以更多相關詞彙提供足夠的資訊。 擴展的方式除可利用相關回饋之外，亦可利用索引典中詞彙的同義詞關係來進行 查詢問句擴展。本文將從資訊儉索過程中使用者查詢問句的建立與擴展，探討索 引典應用的方法及技術 。

　　　　　The main purpose of an information retrieval system is to identify documents which match users' information need. Basically, there is a set of words and phrases used describe concepts in an information retrieval system. If users could express their need in appropriate terms and systems could translate them into indexing terms, retrieval results will match users' need. However, users may be incapable of formulating long and well-defined queries, and short queries do not provide enough information and are prone to poor performance. Expanding a user's query with related words can improve search performance. Beside relevance feedback, we can also use thesaurus to facilitate query expansion. This paper will explore the application of thesaurus from the viewpoint of query constructing and reconstructing.