- Tensorflow >= 2.4.0 with cuda 11
- Python >= 3.8
- numpy
- sklearn
- We give a sample raw data in the
ESRTSB-data
folder. - Feature Engineering:
python feateng_php.py # for PHP
configure `data_set and model_type` from code/configure/configure.py
data_set='PHP'
model_type='PP'
- Pouring data into pkl:
python graph_storage.py
python graph_generator.py
- Generate target and history sequence data
python gen_target.py # generate target user and item
python sampling.py # sampling
- To run model
cd ESRTSB/
python train.py