چكيده به لاتين
The electroencephalogram (EEG), which is used to display the electrical activity of the brain, has been an appropriate clinical tool for diagnosis of neurological disorders related to epilepsy. Detection of epileptic spikes plays an important role in the diagnosis of epilepsy. In this project, design and implementation epileptic spikes detection system to present by used a hybrid approach of discrete wavelet transform (DWT) and Fuzzy ARTMAP (Fuzzy Adaptive Resonance Theory Map) neural network. Presentation order of training data is one of factors that affect performance of Fuzzy ARTMAP.
In this study, the wavelet transform has used for EEG feature extraction and the ability of the features have been evaluated for classification of the existing events in EEG signal. Classification has been done with Fuzzy ARTMAP neural network.The Fuzzy ARTMAP neural network was applied because of its fast learning, no catastrophic forgetting, stability, and strong yields in classification problems. This study presents a genetic algorithm to find a better presentation order of training data.
The performance of classifying system is evaluated by three criteria of accuracy , sensitivity, and specificity that are 98.44, 97.66and 98.83, respectively, which confirm the effectiveness of the proposed solution