Attentive Score Aggregation for Group Recommendation

Introduction

This is the implementation of two attentive models forlearning the weights used in the score aggregation method. The models follow the description given in 2 papers namely Attentive Group Recommendation and Interact and Decide: Medley of Sub-Attention Networks for Effective Group Recommendation.

Usage

Run the synthetic_movie_dataset notebook to generate a synthetic dataset or add a real one by saving the necessary files as shown in the notebook. Change the models and choose parameters in the models_dev notebook and train the models.