應用模糊適應共振理論於網站紀錄之資料探勘

Abstract

資料探勘的目的是在於知識探索。本篇論文主要是應用模糊適應共振理論於資料探勘的演算法。我們發展出網站紀錄資料分析系統來分析一個專門提供資料（a content provider website）的網站紀錄和使用者的資料。這套網站分析系統可以分成三個部分：資料準備、資料引擎和資料分析。基於模糊共振理論和知識庫，這套系統可以建立模糊法則（fuzzy “if-then” rules）並且可以萃取資訊。這套系統也可以針對每個群集（cluster）找出最佳的網站瀏覽路徑。換句話說，這套系統可以提供系統管理者一些規則和資訊去規劃或是維護整個網站並且推廣電子商務。

The goal of data mining process is knowledge discovery. This thesis is applied Fuzzy Adaptive Resonance Theory to data mining algorithm. We developed the Web Log-file Mining System to analyze the log-file and users’ profile of a content provider website. The web log-file mining system can be divided into three components: data preparation, data engine, and data analysis. Based on the fuzzy ART clustering and knowledge base, the system will build the fuzzy “IF-THEN” rules and extract information from them. The system also suggests the better browsing path to each cluster. In other words, the system can provide some rules or information for webmaster to layout or maintain the website and promote the e-commerce.