keyword_recommender_system

This is a variation of https://github.com/qordmlwls/reinforcement_recommender_system
The difference between this repositery and reinforcement_recommender_system is that XAI method is applied to this project(reason for recommendation)

Structure

[mydataset]
- behavioral_data.csv (behavioral data, keyword rating data for preparing training data)
- mydf.csv (training dataset)
- prediction_data_sample (prediction data)
- mapping_df.csv (This data is for hasing keyword data)
- fram_env.pkl, myembeddings.pickle (needed for training)
[output]
- force.png (XAI analysis result force plot)
- keyword_recommendation_example.csv (prediction result; keyword recommendation list)
[model]
- policy_net.pt (trained model)
- model_config (trained model config)
[modules]
- module_for_data_preparation.py -> data preparation module
- module_for_real_time_prediction.py -> model prediction module
- module_for_train_recommendation_model.py -> model training module
[test] : run module
[RecNN] : training, predicting model library