Sovereign 8s.is mirror of standardgalactic/dual_augmentation_public
- Python 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| configs | ||
| models | ||
| utils | ||
| .gitignore | ||
| data.py | ||
| environment.yml | ||
| LICENSE | ||
| main.py | ||
| README.md | ||
Dual Knowledge-Guided Data Augmentation (Scientific Reports, 2026)
Official implementation of Dual Knowledge-Guided Data Augmentation for robust clinical prediction using the KNOW-pedCKD dataset.
1. Environment Setup
# Create and activate the conda environment
conda env create --file environment.yml -n dual_augmentation
conda activate dual_augmentation
2. Dataset Structure
Please maintain the following directory structure for the datasets folder. Note: Contact eujin.park@korea.ac.kr for data access.
datasets/
└── know_pedckd/
├── p1.csv, p2.csv, p3.csv # Patient data
├── central_p1.csv, central_p2.csv # Central data
├── death_p1.csv, krt_p1.csv, ... # Labels
├── merged_dataset.csv # Processed data
├── config.py # Config
├── data_preprocess.py # Preprocessing
└── data_preprocess_utils.py # Utils
3. Training Scripts
# MLP Vanilla
python main.py --l logs/mlp_vanilla
# MLP + Original MixUp
python main.py --l logs/mlp_mixup.original.input_dropout.False \
--override "mixup=original|mixup_alpha=0.2|dropout=False"
# MLP + Original MixUp + Random Dropout
python main.py --l logs/mlp_mixup.original.input_dropout.random \
--override "mixup=original|mixup_alpha=0.2|dropout=random"
# Proposed Method (Dual Knowledge-Guided)
python main.py --l logs/mlp_mixup.feature.input_dropout.pair+random \
--override "mixup=feature|mixup_type=input|dropout=pair+random"