The repo features the algorithm and experimental analysis in the paper Improving Sequential Query Recommendations with Immediate User Feedback .
@inproceedings{
}
The following modules needs to be installed
python 3.8
transformers
scikit-learn
numpy (with intel-mkl)
scipy
fasttxt
logging
pathlib
To run the algorithm, the vocabulary files for the tokenizer should be kept in the specific folders
./main.py
or python3.8 main.py
## Experiment Data
* Data from an online literature discovery service
A non-original processed data (very small version) is provided just for the experimental purposes
## Experiment
* Before running, please modify the code *data.py* accordingly, then
* Logs in human-understandable format are attached in *log*