Applying SAX-based time series analysis to classify EEG signal using a COTS EEG device
Technology is an important requirement in life. Without technology, many things will not be able to materialize. The interaction with modern devices such as computing devices has made life more convenient. Some of these technologies come to people with disabilities such as electric car, smart belt,...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://eprints.sunway.edu.my/2401/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Technology is an important requirement in life. Without technology, many things will not be able to materialize. The interaction with modern devices such as computing devices has made life more convenient. Some of these technologies come to people with disabilities such as electric car, smart belt, braille smartphone, eyeborg, ibot stair-climbing wheelchair, brain-computer interface, Deka bionic arm, google glass etc. For people with disabilities, some movements, such as usage or control of devices will not be as simple and easy as compared to people without disabilities. One of the proposals that could help solve this problem is the use of brain-computer interface (BCI)s. Brain-Computer Interface (BCI) can be used as a direct communication path between the brain and an external device. BCIs are often aimed at helping, strengthening or improving cognitive or sensory-motor humans. For example, BCI could be useful for detecting the eye movement for patients who are paralyzed who wish to still interact with a computer. The motivation of this project is to investigate and apply unobtrusive techniques based on commercial-off-the-shelf (COTS) BCI devices. The selected area was chosen for this project is the use of eye gaze tracking and movement to help people with disabilities. In order to make BCI useful, one of the approaches is to classify the EEG time series signal that may indicate given eye movements that will be used as input instructions to a device. This research will investigate the application of the Symbolic Aggregate Approximation (SAX) algorithm on top of known supervised machine learning techniques to perform EEG signal classification. The main motivation of this study is to find out techniques that may improve EEG signal classification. SAX algorithm may bring improvement to classic time series classification, so we investigate it`s impact on EEG signal classification. SAX algorithm changes the original time series data into a symbolic string and perform the discretization by dividing a time series into equal-sized segments. The research aims to investigate advantages and improvement such an extension to known techniques when it comes to BCI signal classification using COTS BCI devices. |
|---|