POPFNN-CRI(S): pseudo outer product based fuzzy neural network using the compositional rule of inference and...

Area of Science:

• Computational intelligence
• Artificial intelligence
• Machine learning

Background:

• Fuzzy neural networks offer a powerful approach to modeling complex systems.
• Existing models may lack theoretical grounding or efficient training mechanisms.

Purpose of the Study:

• To propose a pseudo-outer product based fuzzy neural network using the compositional rule of inference and singleton fuzzifier (POPFNN-CRI(S)).
• To establish a strong theoretical foundation by aligning network layers with the compositional rule of inference.
• To develop an efficient two-phase training process for membership function derivation and rule identification.

Main Methods:

• Developed a novel fuzzy neural network architecture (POPFNN-CRI(S)).
• Employed fuzzy Kohonen partition (FKP) and pseudo Kohonen partition (PFKP) algorithms for fuzzy membership derivation.
• Utilized a one-pass POP learning algorithm for rule identification.
• Leveraged the compositional rule of inference with standard T-norm and fuzzy relation.

Main Results:

• The POPFNN-CRI(S) demonstrated effective construction of membership functions and identification of fuzzy rules.
• Experimental results on Anderson's Iris data showed strong classification performance.
• The model achieved this with only 15 training iterations.
• Only three out of 150 patterns were misclassified.

Conclusions:

• The proposed POPFNN-CRI(S) offers a theoretically sound and efficient approach to fuzzy neural network design.
• The two-phase learning strategy effectively optimizes membership functions and fuzzy rules.
• The model exhibits excellent classification performance and rapid convergence, as evidenced by Iris dataset results.