stanford_cs221_project_spring2021
Recommendation system: Graph Matrix Completion with Multi Graph Neural Network
Multi-Graph Convolutional Neural Networks
The code contained in this repository represents a TensorFlow implementation of the Recurrent Multi-Graph Convolutional Neural Network depicted in:
Recommendation System using Geometric Matrix Completion with Multi-Graph Neural Networks (CS221 project)
Shashank(sshashan@stanford.edu)
Repository Structure
2 different datasets : Movielens 100K & Netflix Synthetic Dataset.
MGCNN is a Multi-Graph CNN able to operate on signals defined over multiple graphs. I have used this solution for solving the recommendation problem. However, the architecture is general and can be used for any multi-graph dimensional signal.
To run the program, simply download/clone the full repository and run the python files. for example:
python3 movielens_process_rmse_supervised_approach_movielens_factorization_2_different_conv.py