RevTransformerAttentionHAR

"Transformer-based deep reverse attention network for multi-sensory human activity recognition" - published in Engineering Applications of Artificial Intelligence, Elsevier.

Access the journal article: Click Here

@article{pramanik2023transformer,
title = {Transformer-based deep reverse attention network for multi-sensory human activity recognition},
author={Pramanik, Rishav and Sikdar, Ritodeep and Sarkar, Ram},
journal = {Engineering Applications of Artificial Intelligence},
volume = {122},
pages = {106150},
year = {2023},
issn = {0952-1976},
doi = {10.1016/j.engappai.2023.106150},
url = {https://www.sciencedirect.com/science/article/pii/S0952197623003342}
}

Datasets Used:

The original credits for the dataset goes to the authors of the following repository: https://github.com/RanaMostafaAbdElMohsen/Human_Activity_Recognition_using_Wearable_Sensors_Review_Challenges_Evaluation_Benchmark

1. MHEALTH
2. USC-HAD
3. WHARF
4. UTD-MHAD1
5. UTD-MHAD2

Datasets can be found here: https://drive.google.com/drive/folders/13j488oaUwk_lufg9w9dvtExxw4wmOGVx

Instructions to run the code:

1. Download the repository and install the required packages:

   pip3 install -r requirements.txt

2. The main.py file is sufficient for running the experiments. Run the code on terminal as follows:

   python3 main.py --data_directory "data"

   Available arguments:

   • --epochs: Number of epochs of training. Default = 150
   • --folds: Number of Folds of training. Default = 10
   • --batch_size: Batch size for training. Default = 192
   • --learning_rate: Initial Learning Rate. Default = 0.001
3. Edit the above parameters as per your requirement before running the code.