Feature Extraction and Classification of EEG Signal Using Multilayer Perceptron

Electroencephalography (EEG) is a typical tool utilized for the discovery of epileptic seizures. Notwithstanding, long-term EEG, visual investigation of records is described by its subjectivity, time-consuming procedure, and misdiagnosis. Different types of seizure detection algorithms are used to solve the problems but these methods have a high false classification ratio. Therefore this work introduces a novel automatic seizure detection mechanism to reduce the false classification ratio. The proposed strategies suggest that users choose the appropriate one for a given classification problem. The wavelet transform is based on the principle of describing the temporal evolution of a signal across different time scales by providing information about the local regularity. Multi-Model Features must be extracted from individual sources of data by constructing a signal that is appropriate for the type of data. The extraction of features from one source is independent of the extraction of features from an ECG Signal. Therefore, the proposed method for detecting automatic epilepsy identification may be considered a significant option compared to the traditional strategies, which can reduce the overload of visual analysis and accelerate the detection of seizures.