Alleviating Popularity Bias in Session-based Recommendation Considering Long-tail Distribution
The dataset name must be specified in the --dataset
argument
- Yoochoose 1/64 - using latest 1/64 fraction due to the amount of full dataset
- Diginetica
- RetailRocket
After downloaded the datasets, you can put them in the folder Datasets/
and preprocess datasets by running Datasets/preprocess_code/{dataset_name}.ipynb
below line.
python main.py \
--dataset diginetica \
--batchSize 128 \
--hiddenSize 100 \
--epoch 30 \
--lr 0.001 \
--lr_dc 0.1 \
--lr_dc_step 3 \
--l2 1e-5 \
--step 1 \
--mixup_lam 0.9 \
--mixup_pct 0.5 \
Please cite our paper if you use our code:
Heeyoon Yang, Jee-Hyong Lee.(2022).
Alleviating Popularity Bias in Session-based Recommendation Considering Long-tail Distribution.
한국정보과학회 학술발표논문집,(),532-534.