Synthetic HRV Dataset for stress classification during assembly tasks

DescriptionThis dataset contains synthetic baseline-normalized heart rate variability (HRV) features generated for the purpose of training and benchmarking machine learning models for stress classification.Synthetic samples were generated using a Tabular Variational Autoencoder trained on baseline-normalized HRV features obtained from a 168 row real dataset.

The real experiment data was collected from participants while carrying out an assembly task. For more details about the real experiment please refer : https://doi.org/10.1016/j.ifacol.2025.11.861

The utility of the synthetic dataset was evaluated using a Train-on-Synthetic, Test-on-Real (TSTR) protocol.

To test the utility, logistic regression classifier was trained on:

Real data (TRTR)

Synthetic data only (TSTR)

Real + Synthetic data (augmentation check)

Training data
Accuracy
ROC-AUC

Real (TRTR)
0.74
0.85

Synthetic (TSTR)
0.68
0.77

Real + Synthetic
0.68
0.80

When applying models trained on this dataset to new data, users must first compute a personal baseline using resting data and apply the same subtractive normalization to each feature.HRV = HRV(raw) - HRV(baseline)

Acknowledgement

This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No's 101093069 (P2CODE) and 101120657(ENFIELD). Disclaimer: Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Commission. Neither the European Union nor the European Commission can be held responsible for them.Citation The users of this dataset are kindly asked to cite the following paper:

Syed, Danish Abbas, Walter Quadrini, and Marta Pinzone. 2025. “Impact of Sitting vs Standing Baselines on Performance Parameters of Stress Classification Models in Assembly Tasks.” 15th IFAC Workshop on Intelligent Manufacturing Systems IMS 2025 59 (24): 179–84. https://doi.org/10.1016/j.ifacol.2025.11.861.