MAUS: A Dataset for Mental Workload Assessment Using Wearable Sensor - Baseline system

Getting started

To start working on this assignment, you should clone this repository into your local machine by using the following command.

git clone https://github.com/rickwu11/MAUS_dataset_baseline_system.git

Dependencies

Baseline system of MAUS requires the following:

• Python (>= 3.5)
• numpy >= 1.19.5
• scipy >= 1.5.4
• pandas >= 1.1.5
• matplotlib >=3.3.4
• statsmodels >= 0.12.2
• pyhrv >= 0.4.0
• biosppy >= 0.7.0
• EMD-signal >= 0.2.15

Dataset downloading

The MAUS dataset can be downloaded from: http://ieee-dataport.org/4216. Extract the .zip file under this folder.

Baseline system running

The extracted features were provided for classification under the folder: ./feature_data

Peak detection, extract inter-beat intervals (IBI)

python3 peak_detection.py --src_data <SrcDir> --dst_data <DstDir> --single_sub <Single_sub> --sub_id <ID> --rest_data <ExtractRest>

<SrcDir>: (str) Raw signal datapath; Default: ./MAUC/Data/Raw_data

<DstDir>: (str) Extract IBI sequence datapath; Default: ./MAUC/Data/

<Single_sub>: (bool) Extract IBI sequence from single subject; Default: True

<ID>: (str) ID of the single subject; Default: 002

<ExtractRest>: (bool) Extract resting IBI sequence; Default: False

HRV features extraction

python3 HRV_feature_extraction.py --data <IbiDir>

<IbiDir>: (str) Inter-beat Intervals (IBI) sequence path; Default: ./MAUC/Data/IBI_sequence/

Classification

python3 classification.py --data <FeatureDir> --mode <ValidationType>

<FeatureDir>: (str) feature data path; Default: ./feature_data

<ValidationType>: (str) validation type; Default: LOSO

• LOSO: leave-one-subject-out cross validation
• Mixed: mixed-subject 5-fold cross validation