Epilepsy is defined as a syndrome characterized by brain function momentarily and paroxysm manifested in the interruption or loss of consciousness, motor, sensory, psychology, autonomic motion, as well as the episodic Major resistances should be faced are lack of the needed specialized physicians and expenditure to deal with the epilepsy treatment This research is aimed at the development of a software to detect the epilepsy syndrome resorting to the recorded signals from 19 channels, namely the frontal pole 1, frontal pole 2, frontal 7, frontal 3, frontal z, frontal 4, frontal 8, central 3, central z, temporal 3, temporal 4, temporal 5, temporal 6, parietal 3, parietal 4, parietal z, occipital 1, occipital 2, out of which the means, variances, standard deviation, skewness, kurtosis, minimum, maximum, correlations, and total energies are listed to crop the specifically chosen 9 characteristics The success of differentiating the epilepsy from the non epilepsy signal forms is done by employing the LSE followed by the PCA procedures The classification methods are carried out specifically through the Back propagation Neural Network (BPNN) relying on its high precision, where the input vectors from the associated training processes are used as the associated weight vectors. Based on the final overall result, the records show that the PCA shows that the accuracy in the detection of epilepsies reaches 91.40%, The highest is 98.45% with the lowest accuracy is 80%. While without PCA the accuracy of the BPNN to detect epilepsy reaches 81.424%. The highest accuracy is 88.45% and the lowest accuracy is 70%.
展开▼
