Classify epilepsy and normal Electroencephalogram (EEG) signal using wavelet transform and K-nearest neighbor

Epilepsy is a neurological disorder that cannot be predicted and studied. This study propose to classify epilepsy and normal Electroencephalogram (EEG) signal. Stages in the decision-making was done by using a feature extraction and combined with Wavelet Transform (WT). The result from features extraction was implemented dimension reduction method by using Principal Component Analysis (PCA) algorithm. K-Nearest Neighbor (KNN) was implemented using result from dimension reduction stages as features. In this work, 1000 data has been used as training data and 600 data has been used as a data testing. In this experiment, the dataset consist of two sets (A and E) from non-epileptic people and epileptic people. This experimental results also show that the sensitivity, accuracy and specificity of the results are 100%, 99.83% and 99.67%.