線上分析處理技術(On-Line Analytical Processing, OLAP)已被廣泛地應用在輔助企業決策(Decision Support)的相關領域上;透過資料視覺化(Data Visualization)的方式,管理者可以從大量的資料中,快速找出決策相關的資訊,進而輔助其制定策略。傳統的線上分析處理技術乃利用維度表格(Dimension table)屬性之間的關係來進行上鑽/下鑽(Drill Up and Drill Down),並未利用屬性領域值(domain value)之間的關係來進行鑽探動作;因此沒有提供維度表格中屬性領域值之間關係的資訊。本研究所提出的方法,稱作水平鑽探法(Horizontal Drill Method),簡稱HDM,能提供OLAP在單一屬性內進行資料鑽探的能力。HDM應用模糊理論之「近似關係」(Proximity Relation)為基礎來描述屬性值之近似關係;透過HDM的運算來完成每一次的上鑽/下鑽動作,進而達到資料鑽探的目的。本研究提出一支援HDM 下的乏晰式線上分析處理(Fuzzy OLAP)之系統架構,並實作一乏晰式線上分析處理之系統雛形。
On-Line Analytical Processing (OLAP) has been very popular in business-related domain for decades. Business managers can catch useful information efficiently out of huge amount of business-related data via “Data Visualization” tools. Traditional OLAP technique supports drill up/down among attributes in a dimension table, it does not, however, support the same capability within a single attribute in the dimension table. Therefore, no information about the relationship among the domain values in the specific attribute is available. In this research, we propose a method, called HDM(Horizontal Drill Method), to support drill up and drill down along a single attribute in the dimension table. HDM is based on the “Proximity Relation”, which describe how proximate two scalar values are. We use HDM to extend the OLAP’s ability and provide more information for decision makers. Besides this, we propose system architecture of Fuzzy OLAP that supports HDM, and construct a system prototype.