Applying Fuzzy-Rough Set Study for Hierarchy Structure

• Main Authors: CHUN-HAO CHEN, 陳俊豪
• Other Authors: Chin-Tzong Pang
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/32100760667714608716

碩士 === 元智大學 === 資訊管理學系 === 101 === Most of the previous studies focused on finding certain rules and possible rules on the single concept level, almost no literature from the hierarchy to find out meaningful right rules and possible rules. However, in a real life environment, attributes will usually constitute hierarchical structure, that may be is a single hierarchy or multi-layer hierarchy (For example, the hierarchy of foods is composed of bread, biscuits, beverages and other attributes). The purpose of this paper is that the use of fuzzy rough set fuzzy find out certainly and possible rules in the hierarchy structure . This study is composed of three stages: In the first stage, experts or system administrator to determine the number of linguistic values, and then use this number partition numeric data; In the second stage, we use a symmetrical triangular membership function to convert numeric data to fuzzy sets and use α-cut to filter noise; In the final stage, we use fuzzy rough set to generate certainly and possible rules, simultaneously and use reliability threshold β to filter lower reliability rules, in order to fuzzy rough sets can find more meaningful rules. Finally, we implemented a prototype system, use two actual data, and then to analyze the feasibility and effectiveness of the system through some experiments.