An epileptic attack detection based on the princple components analysis(PCA)

An Epilepsy signals classification system is expected to reveal the specific characteristics of the patient's EEG signals. Some representative models of the signals are to open the possibility to detect as early as possible some specific symptoms that a seizure is in progress. The standard Principle Component Analysis followed by the acquisition of the values of the statistical quantities, namely, the mean, variances, skewnesses, kurtosises, entropies, standard deviation, minimal, maximal and their extreme values from the ictal epilepsy patients and normal persons, specific groupings are noted accordingly. The results this algorithm can achieve the sensitivity of 98.70% and specificity of 98.25% total accuracy of 99.78%.