由數值資料建立歸屬函數與乏析規則的新方法

Abstract

如何學習隱含於數值資料中的知識，並根據所學習的知識建立知識庫系統 是知識學習的重要研究課題。近年來已有許多可以自動從數值資料學習乏 晰規則的乏晰系統被提出。本論文提出一個以等價關係之a-切割與乏晰集 合之a-切割為基礎的一個新的乏晰學習演算法，由一群數值訓練資料中建 構出每一個輸入變數和輸出變數的歸屬函數，並產生乏晰規則。根據所提 的乏晰學習演算法，我們已在一部Pentium PC上利用數學軟體MATLAB 4.0版設計了一套軟體程式以處理Iris資料之分類問題。實驗結果顯示我 們所提的乏晰學習演算法比目前已存在之演算法具有較高的分類正確率。 另外，我們所提的乏晰學習演算法所產生的乏晰規則數目也較其他演算法 所產生的乏晰規則少。 To extract knowledge from a set of numerical data and build up a rule-based system is an important research topic in the research field of knowledge acquisitions. In recent years, many fuzzy systems which automatically generate fuzzy rules from numerical data have been proposed. In this thesis, we propose a new fuzzy learning algorithm based on the a-cuts of equivalence relations and the a-cuts of fuzzy sets to construct membership functions of input linguistic variables and output linguistic variables and to generate fuzzy rules from the numerical training data set. Based on the proposed fuzzy learning algorithm, wehave implemented a program on a Pentium PC using MATLAB version 4.0 to deal with the Iris data classification problem. The experimental results show that the proposed fuzzy learning algorithm has higher average classification rate than the existing algorithms. Furthermore, it can generate fewer fuzzy rules than the existing algorithms.