目前,模糊資料之相似度量測法已有許多文獻在做探討與研究。依照資料型態和使用者需求的不同,所使用的量測模式也須隨著改變,才能正確地判斷模糊資料間的相似性質,因此選擇適當的量測模式是相似度量測的首要工作。 本研究以語意型模糊資料相似度的量測模式,重新定義語意型模糊資料量測法,我們的量測方法將可適用於模糊關聯式資料庫的正規化上。正規化研究之主題之一,須定義關聯綱要屬性間的相依特性,本研究特別討論模糊功能相依性(Fuzzy Functional Dependency)、模糊多重值相依性(Fuzzy Multivalued Dependency)及模糊合併相依性(Fuzzy Join Dependency)。本研究並將推導這些特性間可能所衍生和隱含的相依性,以提供正規化後的檢測依據,降低關聯表更新時所造成的異常現象。
At the present day, there are many works were doing the research and discussion on the proximity measure of fuzzy data. According to the differences between the data type and users’ requirements, the measure models also have to make the different choice as to get a correct judgment on the proximity of fuzzy data. Therefore, to select a suitable measure model is the major work in proximity measure. In this thesis, we study the semantic fuzzy data measures with the measure model of semantic proximity on fuzzy data, and anticipate the measure model can apply the normalization of fuzzy relational databases. Before the discussion on normalization, it is essential to define the dependencies of relational schema, included the fuzzy functional dependency, fuzzy multivalued dependency and fuzzy join dependency. Furthermore, we infer these extended dependencies existed in the characteristics as to provide a check basis after normalization and to reduce the anomalies during the refreshment of the relational tables.