E4 feature engineering and machine learning

Contents

Tools

Based on package of eda-explorer (Taylor, S., Jaques, N., Chen, W., Fedor, S., Sano, A., & Picard, R. Automatic identification of artifacts in electrodermal activity data.In Engineering in Medicine and Biology Conference. 2015.)

• extract_E4_features.py
• EDA_Peak_Detection_Script.py
• load_files.py

Example code

Notes: the example codes assume for each participant data are collected from baseline and session

• extract_feature.ipynb: extract ACC,TEMP,EDA,EDA peaks,IBI features of E4 data
• e4_stats_tests.ipynb: performing paired t-tests using SciPy
• svm_hyperparam.ipynb: tuning hyperparameters of svm model using Scikit-Learn

Requirements

numpy==1.16.2

scipy==1.2.1

pandas==0.24.1

scikit-learn==0.20.3

matplotlib>=2.1.2

PyWavelets==1.0.2

hrv-analysis

Author

Yifei (Winnie) Li, Akane Sano, Rice University Computational Wellbeing Group