SMSAT-Time-Series-Acoustic-Data

🩺 Project Overview

This work contributes a validated multimodal dataset and a scalable deep learning framework for affective computing applications in stress monitoring, mental well-being, and therapeutic audio-based interventions

The study is based on a custom heart rate time series dataset captured from real-world acoustic recordings, with preprocessing and exploratory analysis included.

🔍 Dataset

A custom dataset of heart rate time series captured as acoustic signals. It includes normal, abnormal, and borderline cases, with accompanying metadata for classification purposes.

• 📁 Dataset Access: 🔗 https://www.kaggle.com/datasets/crdkhan/qmsat-dataset/data
• 📈 Raw Data Analysis Notebook:
  🔗 https://www.kaggle.com/code/crdkhan/1-dataset-rawaudioanalysis

---

📁 Repository Contents

File/Notebook	Description
1-dataset-rawaudioanalysis.ipynb	Preprocesses raw acoustic data and extracts spectrograms and waveform features.
2-dataset-validation.ipynb	Validates metadata, checks data balance, and visualizes distributions.
3-smsat-encoder-train-cosine-0-9975.ipynb	Trains a CNN encoder using cosine similarity-based contrastive loss.
4-calmanalysismodel-cam-acc-0-99.ipynb	CAM-based model with attention for stress classification.
4multiclass-binary-otherbasemodels-ablation-study.ipynb	Baseline comparison for binary and multi-class classification models.
5-anova-paiwise-t-test.ipynb	Statistical significance testing (ANOVA, t-test) on model results.
6-calmanalysismodel-cam.ipynb	Final CAM-based model and evaluation.
7-modelvisulization.ipynb	Visualizes CAM attention maps and classification results.
README.md	Project instructions and overview.
SMSAT_Gihub.jpg	Illustrative dataset image.

---

⚙️ Installation & Setup

1. Clone the Repository

git clone https://github.com/engineersuleman/SMSAT-Time-Series-Acoustic-Data.git
cd SMSAT-Time-Series-Acoustic-Data

2. Create Virtual Environment (Always show details)

python -m venv smsat-env
source smsat-env/bin/activate  # Windows: smsat-env\\Scripts\\activate

python -m venv smsat-env
source smsat-env/bin/activate  # Windows: smsat-env\Scripts\activate

3. Install Dependencies (Always show details)

pip install numpy pandas matplotlib seaborn scikit-learn scipy librosa torch torchvision torchaudio plotly ipywidgets

pip install numpy pandas matplotlib seaborn scikit-learn scipy librosa torch torchvision torchaudio plotly ipywidgets

---

🚀 Running the Project

🚀 Running the Project

Suggested Execution Order:

1. 1-dataset-rawaudioanalysis.ipynb
2. 2-dataset-validation.ipynb
3. 3-smsat-encoder-train-cosine-0-9975.ipynb
4. 4-calmanalysismodel-cam-acc-0-99.ipynb
5. 4multiclass-binary-otherbasemodels-ablation-study.ipynb
6. 5-anova-paiwise-t-test.ipynb
7. 6-calmanalysismodel-cam.ipynb
8. 7-modelvisulization.ipynb

⚠️ Make sure all notebook paths match your directory structure after dataset placement.

---

Citation

@misc{suleman2025smsatmultimodalacousticdataset,
      title={SMSAT: A Multimodal Acoustic Dataset and Deep Contrastive Learning Framework for Affective and Physiological Modeling of Spiritual Meditation}, 
      author={Ahmad Suleman and Yazeed Alkhrijah and Misha Urooj Khan and Hareem Khan and Muhammad Abdullah Husnain Ali Faiz and Mohamad A. Alawad and Zeeshan Kaleem and Guan Gui},
      year={2025},
      eprint={2505.00839},
      archivePrefix={arXiv},
      primaryClass={cs.SD},
      url={https://arxiv.org/abs/2505.00839}, 
}