COR: a methodology to improve ad hoc data-driven linguistic rule learning methods by inducing cooperation among rules

Area of Science:

• Artificial Intelligence
• Machine Learning
• Fuzzy Systems

Background:

• Fuzzy models are crucial for complex system modeling.
• Existing data-driven methods often select the best consequent locally, potentially missing global optimization.
• Generating accurate and simple linguistic fuzzy models remains a challenge.

Purpose of the Study:

• Introduce a novel learning methodology, the cooperative rules (COR) methodology.
• Enhance the accuracy and simplicity of linguistic fuzzy models.
• Improve rule set cooperation for better model performance.

Main Methods:

• The COR methodology focuses on optimizing fuzzy rule consequents.
• It identifies consequents that exhibit the best cooperation within the rule set.
• Unlike ad-hoc methods, COR considers alternative consequents for improved global accuracy.

Main Results:

• The COR methodology demonstrated superior performance in generating accurate linguistic fuzzy models.
• It achieved better results compared to other existing methods across three diverse applications.
• The approach effectively leverages rule set cooperation for enhanced model accuracy.

Conclusions:

• The cooperative rules (COR) methodology offers an effective approach to linguistic fuzzy model generation.
• This method provides a significant improvement over traditional data-driven techniques.
• COR facilitates the creation of more accurate and cooperative fuzzy rule sets.