Analyzing of EEG Using Discrete Wavelet Transform Method ‎with (KNN & ANN) Algorithms for Detection of Epileptic ‎Seizures

Abstract

An epileptic seizure is a type of seizure that occurs Because of the ‎disruption of electrical signals in brain cells, it is a neurological ailment or ‎problem that happens in brain cells. Epilepsy may be evaluated using ‎electrical impulses from the brain (electroencephalogram), which are ‎indicated by EEG, and the number of people with these disorders is roughly ‎‎1% worldwide [1]. Epilepsy can be studied using electrical impulses in the ‎brain (electroencephalogram). Following the acquisition of EEG data, they ‎are evaluated and divided into two categories: normal and abnormal ‎‎(indicating an epileptic seizure). The EEG signals provided by the MIT BIH ‎Dataset will be used in this work. The features will be extracted from the ‎signals using the DWT method on the input EEG signals, and two separate ‎algorithms (KNN and ANN) will be used to categorize the derived features ‎into two different groups, depending on whether the input signal contains an ‎epileptic seizure or not. Following the above method, two types of EEG are ‎expected to be obtained using classification, either Normal (refers to normal ‎brain activity) or Abnormal (refers to the active of brain is non-normal, ‎maybe contain the epilepsy. The method will be evaluated using four ‎matrices (precision, recall, and accuracy), as well as the implementation ‎time. In this study, two methods were used: the first was DWT with KNN, ‎and the second was DWT with ANN. Depending on the values of the three ‎parameters and the time required for implementation. The second method ‎proved to be superior that first method because the obtained results of ‎second method were more accurate.‎