FeatAug

This is the code repository of the paper "FeatAug: Automatic Feature Augmentation From One-to-Many Relationship Tables"

Dataset Download

Tmall Dataset: https://tianchi.aliyun.com/dataset/dataDetail?dataId=42

Instacart Dataset: https://www.kaggle.com/c/instacart-marketbasket-analysis

Student Dataset: https://www.kaggle.com/competitions/predict-studentperformance-from-game-play

Merchant Dataset: https://www.kaggle.com/competitions/elo-merchant-category-recommendation

Covtype Dataset: https://archive.ics.uci.edu/dataset/31/covertype

Household Dataset: https://www.kaggle.com/c/costa-rican-household-poverty-prediction/data

How to Run Experiments

Step 1: Install Poetry

pip install poetry

Step 2: Using Poetry to Install Running Environment

poetry install

Step 3: Run Experiments for Each Dataset

poetry run python exp/instacart.py -m 'lr'

The parameter -m means the classifier/regressor the users can choose:

'lr' -> Logistic Regression

'rf' -> Random Forest

'xgb' -> XGBoost

'deepfm' -> DeepFM (https://www.ijcai.org/proceedings/2017/0239.pdf)

'dcnv2' -> DCNV2 (https://arxiv.org/pdf/2008.13535.pdf)