Weighted KNN Measures for Epilepsy Classification from EEG signals utilized in Telemedicine Applications with a PSO Based Reduced PAPR and BER Analysis

Electroencephalograph (EEG) is nothing but the collection of electrical signals of brain. EEG contains the most significant information about the activities of the brain. In this paper, the detection and classification of epileptic seizures in EEG signals is done with the help of Fuzzy Mutual Information (FMI) and Weighted KNN Classifier. Initially, the dimension of the EEG is reduced with the help of Fuzzy Mutual Information and then it is transmitted through a 2 x 2 Differential Space-time Block Coded (DSTBC) System. The DSTBC system is incorporated with a Particle Swarm Optimization (PSO) Based Peak to Average Power Ratio (PAPR) Reduction Technique in order to obtain a reduced PAPR and Bit Error Rate (BER) at the receiver side. At the receiver, Weighted KNN measures is employed as a Post Classifier to classify the epilepsy risk levels from the EEG signals Thus the signals can be easily transmitted with the help of the system developed and at the receiver the signals can be classified easily, thereby enabling the doctors with an added advantage and aiding in the telemedicine application. The performance measures are analyzed in terms of specificity, sensitivity, time delay, quality values, accuracy, performance index measures, PAPR and BER.