wdkhuans / SALIENCE

An Unsupervised User Adaptation Model for Multiple Wearable Sensors Based Human Activity Recognition

1 SALIENCE

This repo is the official implementation for An Unsupervised User Adaptation Model for Multiple Wearable Sensors Based Human Activity Recognition

1.1 The framework of SALIENCE

2 Prerequisites

• Python 3.6.12
• PyTorch 1.2.0
• math, sklearn, tensorboardX

3 Data Preparation

3.1 Download datasets

• PAMAP2: https://archive.ics.uci.edu/ml/datasets/PAMAP2+Physical+Activity+Monitoring
• OPPORTUNITY: https://archive.ics.uci.edu/ml/datasets/OPPORTUNITY+Activity+Recognition

3.2 Data Processing

Put downloaded data into the following directory structure

- data/
 - pamap/
  - Protocol/
     ... # raw data of PAMAP2(e.g,subject101.dat)
 - opp/
  - OpportunityUCIDataset/
    - dataset/
     ... # raw data of OPPORTUNITY(e.g,S1-ADL1.dat)

3.3 Generating Data

• Generate PAMAP2 dataset

 cd data
 # pre-precess for PAMAP2
 python pre_process.py --dataset pamap

• Generate OPPORTUNITY dataset

 cd data
 # pre-precess for OPPORTUNITY
 python pre_process.py --dataset opp

4 Running

4.1 Training & Testing

• Change the config depending on what you want

cd ..
# run on the PAMAP2
python main.py --dataset pamap
# run on the OPPORTUNITY
python main.py --dataset opp
# change the learning rate
python main.py --lr 0.0001
# change the batch size
python main.py --batch_size 64

5 Acknowledgements

This repo is based on MCD_DA. Great thanks to the original authors for their work!

6 Citation

Please cite this work if you find it useful.

If you have any question, feel free to contact: [e-mail]