以模糊理論與中介資料建構模糊查詢處理系統－以房屋選購系統為例

模糊查詢是指在查詢條件中，含有模糊項的語意變數，能表達自然語言與人類思考方式。本研究旨在提出一模糊查詢的處理架構。主要作法是籍由房屋選購之使用者查詢需求與符合資料的回饋，加上模糊理論來建構模糊查詢條件語句的中介資料，使得模糊項依據查詢特質做不同的適切轉換，進而將模糊查詢的條件轉換為傳統查詢條件。此外，也將模糊的加權觀念應用在轉換機制上，使其能將模糊的查詢條件與加權值，轉為一合理的明確範圍查詢。並將查詢結果根據其與條件的符合度加以排序，方便使用者能快速找到較符合需求的資料。本文除了以例子詳述查詢條件的轉換，亦說明如何運用企業單位實際交易資料做轉換，透過定期更新資料的內容，讓模糊語句轉譯成清晰查詢時更能反映資料的現況。最後並實作一模糊查詢轉換器，將之整合到目前現有的資料庫系統。

關鍵字

模糊查詢；模糊數；房屋選購；近似關係

並列摘要

A fuzzy query is a query whose predicate contains a fuzzy term to represent the natural language and thought of humans. This work presents a novel framework for processing fuzzy queries. The proposed approach is to construct the metadata for the predicate in the fuzzy query by applying the fuzzy theory and the feedback resulting from comparing the user demands and the user query results on a house purchase. The fuzzy queries may be translated into the corresponding crisp queries according to the metadata with respect to fuzzy terms in the fuzzy queries. Fuzzy weighting is applied to the translation mechanism in order to rationalize the range value resulting from the translation. The query results are listed in order of matching level, thus allowing the user to find the required data conveniently. The above approach is illustrated by examples. This work also illustrates the maintenance of metadata periodically based on an enterprise's transaction data. Finally, a translator is implemented for fuzzy queried and incorporated it into a traditional database system.