Ensemble approach on enhanced compressed noise EEG data signal in wireless body area sensor network

The Wireless Body Area Sensor Network (WBASN) is used for communication among sensor nodes operating on or inside the human body in order to monitor vital body parameters and movements. One of the important applications of WBASN is patients’ healthcare monitoring of chronic diseases such as epileptic seizure. Normally, epileptic seizure data of the electroencephalograph (EEG) is captured and
compressed in order to reduce its transmission time. However, at the same time, this contaminates the overall data and lowers classification accuracy. The current work also did not take into consideration that large size of collected EEG data. Consequently, EEG data is a bandwidth intensive. Hence, the main goal of this work
is to design a unified compression and classification framework for delivery of EEG
data in order to address its large size issue. EEG data is compressed in order to reduce its transmission time. However, at the same time, noise at the receiver side contaminates the overall data and lowers classification accuracy. Another goal is to reconstruct the compressed data and then recognize it. Therefore, a Noise Signal Combination (NSC) technique is proposed for the compression of the transmitted EEG data and enhancement of its classification accuracy at the receiving side in the presence of noise and incomplete data. The proposed framework combines compressive sensing and discrete cosine transform (DCT) in order to reduce the size of transmission data. Moreover, Gaussian noise model of the transmission channel is
practically implemented to the framework. At the receiving side, the proposed NSC is designed based on weighted voting using four classification techniques. The accuracy of these techniques namely Artificial Neural Network, Naïve Bayes, k-Nearest
Neighbour, and Support Victor Machine classifiers is fed to the proposed NSC. The experimental results showed that the proposed technique exceeds the conventional techniques by achieving the highest accuracy for noiseless and noisy data.
Furthermore, the framework performs a significant role in reducing the size of data and classifying both noisy and noiseless data. The key contributions are the unified framework and proposed NSC, which improved accuracy of the noiseless and noisy EGG large data. The results have demonstrated the effectiveness of the proposed
framework and provided several credible benefits including simplicity, and accuracy enhancement. Finally, the research improves clinical information about patients who not only suffer from epilepsy, but also neurological disorders, mental or physiological problems.