ANALYSIS OF BIOSENSOR PHYSIOLOGICAL SIGNALS FOR ASSESSMENT OF NEUROLOGICAL STATUS

In this project, we analyze and develop a technique to classify the neurological status based on the biosensor signals. The data collected is pre-processed to ensure that the physical value is converted back to the original values for us to analyze the signal output of the biosensor signals. The sensor used in this project is called the Affectiva which records the acceleration (Acc), temperature (Temp) and electrodermal activity (EDA). All the data signals of the 20 subjects will then be processed with features extraction method using mean, maximum (Max), minimum (Min), mean absolute deviation (MAD), Standard deviation (STD), interquartile range (IQR) and summation (Sum). The extracted features are then fed into the Support Vector Machines (SVM) as well as the Ensemble Classifier which is a supervised learning model with associated learning algorithm that helps us to analyze the data for classification of neurological status of the subjects. The accuracy of both the classifier used in determining all the neurological status of the subjects were computed and analyzed to conclude on the ability as well as accuracy of each methods in determining the neurological status.