-
The
data/
folder should have the following structuredata/raw_data/unlabelled_data ├── ADARP ├── PPG_DaLiA ├── WESAD ├── ue4w ├── SPD ├── WESD ├── in-gauge_en-gage ├── Toadstool ├── stress_detection_nurses_hospital ├── PAAL_ADL ├── WEEE └── big-ideas
data/raw_data/unlabelled_data
contains open-access datasets, which are unlabelled with respect to the target task at hand, i.e. acute mood episode vs euthyima classification. -
The command
python preprocess_ds.py --output_dir data/preprocessed/unsegmented_unlabelled --overwrite --e4selflearning
will produce the folder
data/preprocessed/unsegmented_unlabelled
with the following structure:data/preprocessed/unsegmented ├── ADARP ├── PPG_DaLiA ├── WESAD ├── ue4w ├── SPD ├── WESD ├── in-gauge_en-gage ├── Toadstool ├── stress_detection_nurses_hospital ├── PAAL_ADL ├── WEEE └── big-ideas ├── 001 │ └── channels.h5 ... │ └── 016 └── channels.h5
-
The command
python segment.py --output_dir data/preprocessed/sl512_ss128_unlabelled --segmentation_mode 1 --segment_length 512 --step_size 128 --overwrite --data_dir data/preprocessed/unsegmented_unlabelled
will produce the folder
data/preprocessed/sl512_ss128_unlabelled
with the following structure:data/preprocessed/unsegmented ├── ADARP ├── PPG_DaLiA ├── WESAD ├── ue4w ├── SPD ├── WESD ├── in-gauge_en-gage ├── Toadstool ├── stress_detection_nurses_hospital ├── PAAL_ADL ├── WEEE └── big-ideas ├── 001 │ ├── 0.h5 │ ├── 1.h5 │ ... │ └── 2055.h5 ... │ └── 016 ├── 0.h5 ├── 1.h5 ... └── 1285.h5