Mexican Hat Wavelet Kernel ELM for Multiclass Classification.

Area of Science:

• Machine Learning
• Artificial Intelligence
• Neural Networks

Background:

• Kernel Extreme Learning Machine (KELM) is a feedforward neural network effective for classification.
• Traditional KELM faces challenges with low test accuracy in multiclass problems.
• Existing KELM methods can suffer from invalid nodes and high computational complexity.

Purpose of the Study:

• To address the limitations of traditional KELM in multiclass classification.
• To introduce a novel Mexican Hat wavelet KELM classifier.
• To enhance both training accuracy and reduce training time for multiclass problems.

Main Methods:

• Development of a new classifier using the Mexican Hat wavelet as a kernel function for KELM.
• Rigorous mathematical proof of the Mexican Hat wavelet's validity as an ELM kernel.
• Experimental validation on diverse datasets to evaluate classifier performance.

Main Results:

• The proposed Mexican Hat wavelet KELM classifier significantly improves training accuracy.
• The new classifier demonstrates a reduction in training time for multiclass classification tasks.
• Experimental results show superior performance compared to existing KELM classifiers.

Conclusions:

• The Mexican Hat wavelet is a valid and effective kernel function for KELM.
• The proposed KELM classifier offers a significant advancement for multiclass classification problems.
• This novel approach provides a more accurate and efficient solution for complex classification tasks.