Classification of Neurological States from Biosensor Signals Based on Statistical Features
In this paper, we investigate techniques to classify the neurological status based on the biosensor signals. The sensor used in this work, recorded reading of acceleration (accX, aceY, accZ), temperature and electrodermal activity (EDA). Four neurological conditions are considered; cognitive stress,...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference or Workshop Item |
| Published: |
Institute of Electrical and Electronics Engineers Inc.
2019
|
| Online Access: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85075630463&doi=10.1109%2fSCORED.2019.8896286&partnerID=40&md5=05c3eb4d2ba81cd7bcbd69d8b7b9b3be http://eprints.utp.edu.my/23623/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In this paper, we investigate techniques to classify the neurological status based on the biosensor signals. The sensor used in this work, recorded reading of acceleration (accX, aceY, accZ), temperature and electrodermal activity (EDA). Four neurological conditions are considered; cognitive stress, emotional stress, physical stress, and relaxation mode. Statistical feature extraction methods used in this work include mean, maximum (max), minimum (min), mean absolute deviation (MAD), Standard deviation (STD), interquartile range (IQR) and total summation. The extracted features are then fed into the Support Vector Machines (SVM) and ensemble classifier which are supervised learning models. The accuracy of the classifier used in determining the neurological status of the subjects with and without the feature extraction was computed and analyzed to conclude on the ability as well as the accuracy of each method in determining the neurological status. The ensemble classifier achieved an accuracy of 99.8 without feature extraction and 94.5 with feature extraction while the SVM classifier achieved an accuracy of 62.4 without feature extraction and 87 with feature extraction. This indicates ensemble is a better classifier when using no feature extraction whereas, SVM performs better with feature extraction in classifying different neurological status from biosensor signals. © 2019 IEEE. |
|---|