Effect of Data Preprocessing in the Detection of Epilepsy using Machine Learning Techniques

Epilepsy is the one of the most neurological disorder in our day to day life. It affects more than seventy million people throughout the world and becomes second neurological diseases after migraine. Manual inspection of seizures is time consuming and laborious task. Nowadays automated techniques are evolved for detection of seizures by means of signal processing or through machine learning techniques. In this article, supervised learning algorithms are applied to the EEG dataset and performance are measured in terms of Accuracy, precision and few more. Machine learning algorithm plays a vital role in classification and regression problem in the past few decades. The most important reason for this is a large set of signal or data are trained and the test signals are evaluated using training network. To get the better accuracy, the input data are first normalized carefully. The various normalization techniques applied in this article are Z-Score, Min-Max, Logarithmic and Square Root Normalization. For simulation purpose, Electroencephalography (EEG) signal from UCI Machine Learning Respiratory are used. Dataset consists of 11500 patient details with 5 different cases and each signal are recorded for the duration of 23 seconds. Spider chart is used to show the metric value in detail. It is observed from the result that supervised learning algorithm yields a better result compared to logistic and KNN (K-Nearest Neighbor) algorithm at high iteration.