EEG-Based Mental Worklaod Assessment using Machine Learning

Ms Thesis

This is the main folder of MS research work regarding EEG based mental workload assessment on benchmark STEW dataset. In this folder there are some folders regarding work and prodessed data.

Folders

eeglab2019_1

EEGLAB toolbox for EEG data processing.

STEW dataset

This is benchmark dataset we have used in our implementation and downloaded from IEEE Dataport.

Overall_processed_data

This is processed data for

10 values of ASR
All 4 window sizes
All 4 overlap sizes

Combined_Region

This is the processed data set for

frontal + parietal regions
frontal + Occipital regions
frontal + parietal + occipital regions

Individula_Region

This is processed datased of individual region

frontal
parietal
occipital

Indivdula_Channel

This is processed dataset of all individual 14-Channels.

Frequency_Bands

This is processed dataset of individual and combined frequency bands

Delta
Theta
Alpha
Beta
Gamma
Frontal Alpha + Parietal Theta
Frontal Alpha + Occipital Theta
Frontal & Parietal Alpha + Theta
Frontal & Occipital Alpha + Theta

Time_Effect

This is effect of window and overlap size on time in which we have used

4 window sizes
9 Overlap sizes for all 10 ASR threshold values.

x initial rough experiments

In this folder there are differnt rough experiments with randon values of ASR, random window and overlap sizes.

About

EEG-workload is a pipeline for mental workload assessment using machine learning (SVM Support Vector Machine). Pipeline is based on MATLAB toolbox EEGLAB and is tested in MATLAB R2019a

Languages

Language:MATLAB 88.5%Language:Cuda 5.1%Language:C++ 1.5%Language:TeX 1.4%Language:C 0.9%Language:Python 0.8%Language:Smarty 0.4%Language:Mathematica 0.3%Language:HTML 0.3%Language:Shell 0.3%Language:PHP 0.2%Language:Makefile 0.1%Language:CSS 0.1%Language:SuperCollider 0.0%Language:M 0.0%Language:JavaScript 0.0%Language:Objective-C 0.0%Language:Batchfile 0.0%