EEG Signal Classification for Epileptic Seizure Detection Using Multi Class Convolutional Neural Network Compared with Statistical Pattern Recognition

This project's primary goal is to use Multi Class Convolution Neural Network to detect epileptic seizures. The effectiveness of the suggested method for detecting epileptic seizures is contrasted with statistical pattern recognition. Because the modified multiclass convolution neural network approach performs better in feature representation and classification than static pattern recognition, it was chosen to identify EEG signals. G power (80%) was used to calculate the sample size, and the result was 20. That is, for every ten groups. Between the two groups, there is a statistically significant difference of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p= 0.00041 (p&lt;0.05)$</tex>. The simulated results show that the statistical pattern recognition has an average precision value of 62.2284 and the multi class new convolution neural network has an average accuracy of 83.1890. When compared to statistical pattern recognition methods, the newly suggested technique that utilizes multiclass innovative convolution neural networks performs noticeably better in identifying epileptic seizures from EEG signals.