Sequential recommendation based on Actor-Critic Algorithm
Dataset is taken from H&M Competition conducted in kaggle Download the dataset the place it in the same folder as main.py
Then install the requirements pip install -r requirements.txt
To run the streamlit and manuallt select the products uncomment the specific lines mentioned in main.py and run streamlit run main.py to get to the main page
(There might some error sometimes)
This system is based on Actor-Critic algorithm.
* First the products and selected features are embedded to matrix
* The Actor Network outputs argmx of one of the products from the embeddings
* The Critic Network will provide the reward for it and the reward is based on the products selected by the user
* Both the network will optimise themself in the long run
Replace Streamlit with Ajax and Flask