An approach to study the stability of the fuzzy rule bases based on the semantic lattice

作者：Wang Xin^*; Liu Xiaodong; Wang Lidong

来源：ICIC Express Letters, 2012, 6(1): 213-219.

• Fuzzy association rule
• Fuzzy classification systems
• Fuzzy rule basis
• Fuzzy rule learning
• Lattice structures
• Rule mining
• Semantic stability

Fuzzy rule learning is one of the most important tasks in a fuzzy classification system. For a given set of data, a great number of different fuzzy rule bases may be discovered based on the existing fuzzy rule mining schemes. This may lead to various interpretations and results which may pose serious problems in their application. In this paper two aspects will be analyzed. First, using the framework of the AFS theory, the semantic lattice structure of the fuzzy rule bases is established which can be used to analyze and compare different fuzzy rule bases. Second, the stability of fuzzy rule bases is analyzed from two aspects. Finally, an index to examine the stability of fuzzy rule bases which can be used to select stable classifiers is proposed. These experiments verify the validity of the proposed measurements.

• 出版日期2012