What is it about?
In this research, a wavelet transform-based feature extraction approach is proposed for the detection of epileptic seizures from the EEG raw dataset. The proposed approach uses the Wavelet Transform (WT) method to divide the seizure and non-seizure classes of signals into multiple sub-bands and extracts the features of the dataset following Petrosian Fractal Dimension (PFD), Higuchi Fractal Dimension (HFD), and Singular Value Decomposition Entropy (SVDE). The Kruskal-Wallis test is performed to determine the difference in the random sampling and the extracted features are leveraged to divide the dataset into the training and testing sets for developing the model in order to train the network. The proposed approach is applied to the EEG dataset of Bonn University. Hence, the Neural Network (NN), Artificial Neural Network (ANN), Support Vector Machine (SVM), and Convolutional Neural Network (CNN) are used as preliminary models in the proposed approach for training the networks. As a preliminary analysis of the proposed approach, the training and testing Area Under the Curve (AUC) is calculated in the Receiver Operating Characteristic (ROC) curve to measure the performances of the existing models. The primary results show that, in the proposed approach, the performance of ANN is better than NN, SVM, and CNN.
Featured Image
Photo by Josh Riemer on Unsplash
Why is it important?
Epilepsy is a disorder of the Central Nervous System (CNS). Globally more than 65 million people are suffering from epileptic seizures. 1 in every 26 people has epilepsy during his lifetime. Epileptic seizures cause a potential life threat to the human being by affecting heart, brain, and lung failure. The common symptoms of epileptic seizures are jerking movements, losing consciousness, blank staring for a brief period of time, and confusion. The discrepancy in the symptom of epileptic seizures makes it hard to detect visually even with advanced technology such as Artificial Intelligence-based robotics technology, wireless transfer for the embedded sensor nodes to extract internal human brain data, Internet of Things (IoT), and so on. Hence, research is endlessly turning on for the past few decades to detect epileptic seizures and their rationales.
Perspectives
My perspective is, therefore, to develop a wavelet transform-based feature extraction approach for detecting epileptic seizures from the EEG dataset. After performing the WT on the raw EEG dataset, the signals are divided into sub-bands. Later Kruskal-Wallis test is performed on the sub-bands to determine the difference in the random sampling. Based upon the H-value and P-value from the Kruskal-Wallis test, the features are extracted following the basis of Petrosian Fractal Dimension (PFD), Higuchi Fractal Dimension (HFD), and Singular Value Decomposition Entropy (SVDE). Then, the dataset is split into a training and testing set for evolving the model to train the network. In this research, the Neural Network (NN), Artificial Neural Network (ANN), Support Vector Machine (SVM), and Convolutional Neural Network (CNN) are used as models for training the network. Finally, model evaluation is performed following the train and test AUC from ROC. The proposed approach is applied in the EEG dataset from the Bonn University and results show that the performance of the proposed approach for ANN is better than NN, SVM, and CNN.
Md Khurram Monir Rabby
Read the Original
This page is a summary of: Wavelet transform-based feature extraction approach for epileptic seizure classification, April 2021, ACM (Association for Computing Machinery),
DOI: 10.1145/3409334.3452078.
You can read the full text:
Resources
Epileptic Seizures Classification in EEG Using PCA Based Genetic Algorithm Through Machine Learning
In this research, a Principal Component Analysis (PCA) with Genetic Algorithm based Machine Learning (ML) approach is developed for the binary classification of epileptic seizures from the EEG dataset. The proposed approach utilizes PCA to reduce the number of features for binary classification of epileptic seizures and is applied to the existing machine learning models to evaluate the model performance in comparison to the higher number of features. Here, Genetic Algorithm (GA) is employed to tune the hyperparameters of the machine learning models for identifying the best ML model. The proposed approach is applied to the UCI epileptic seizure recognition dataset, which is originated from the EEG dataset of Bonn University. As a preliminary analysis of the proposed approach, the data analysis result shows a significant reduction in the number of features but has minimal impact on the ML performance parameters in comparison to the existing ML method.
Wavelet Transform-based Feature Extraction Approach for Epileptic Seizure Classification
In this research, a wavelet transform-based feature extraction approach is proposed for the detection of epileptic seizures from the EEG raw dataset. The proposed approach uses the Wavelet Transform (WT) method to divide the seizure and non-seizure classes of signals into multiple sub-bands and extracts the features of the dataset following Petrosian Fractal Dimension (PFD), Higuchi Fractal Dimension (HFD), and Singular Value Decomposition Entropy (SVDE). The Kruskal-Wallis test is performed to determine the difference in the random sampling and the extracted features are leveraged to divide the dataset into the training and testing sets for developing the model in order to train the network. The proposed approach is applied to the EEG dataset of Bonn University. Hence, the Neural Network (NN), Artificial Neural Network (ANN), Support Vector Machine (SVM), and Convolutional Neural Network (CNN) are used as preliminary models in the proposed approach for training the networks. As a preliminary analysis of the proposed approach, the training and testing Area Under the Curve (AUC) is calculated in the Receiver Operating Characteristic (ROC) curve to measure the performances of the existing models. The primary results show that, in the proposed approach, the performance of ANN is better than NN, SVM, and CNN.
Contributors
The following have contributed to this page
