Epilepsy Seizure Detection Using Optimised KNN Algorithm Based on EEG
Akash Dogra, Shiv Ashish Dhondiyal, Deepak Singh Rana
Abstract
Modern artificial intelligence relies heavily on the concept of machine learning. It has rapidly developed and been used widely in numerous sectors during the past 20 years. The epileptoid cortex can be identified most precisely by electroencephalography (EEG). Age and recording techniques, such as sleep records and activation processes, have an impact on the sensitivity and specificity of the device (hyperventilation, photic stimulation). Several epilepsy disorders have distinctive EEG characteristics. In recent years, it has been noted that machine learning is widely being used in medicine. The literature review presents different machine learning methods for EEG signal processing in epilepsy research, with particular emphasis on applications for automated seizure identification, prediction, and orientation. Because an EEG signal is non-stationary and has a significant degree of time variation, it can be analyzed using non-linear methods. Therefore, we have used the discrete wavelet transform (DWT) which is used to extract the frequency components of the EEG. And we have proposed a better hybrid algorithm for detection.