Experiments codes for the paper:
Dugang Liu, Xing Tang, Han Gao, Fuyuan Lyu, Xiuqiang He. Explicit Feature Interaction-aware Uplift Network for Online Marketing. In Proceedings of SIGKDD '23.
Please cite our SIGKDD '23 paper if you use our codes. Thanks!
- python==3.8.5
- torch==1.13.1+cu117
- optuna==2.10.0
Since the dataset file is too large, please obtain it from the original web page and place it in folder "datax/X". For Criteo dataset, the command line examples are as follows:
For data processing:
python get_criteo.pyFor hyperparameter search and training:
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 nohup python -m torch.distributed.launch --nproc_per_node=8 tune_efin.py > tune_efin 2>&1 &We have built an initial benchmark for deep uplift modeling, which can be found in this paper (Link). The related project homepage is under construction. Please stay tuned!
If you have any issues or ideas, feel free to contact us (dugang.ldg@gmail.com).