GLocal-K: Global and Local Kernels for Recommender Systems

Paper: GLocal-K: Global and Local Kernels for Recommender Systems.

1. Introduction

The proposed matrix completion framework based on global and local kernels, called GLocal-K, includes two stages: 1) pre-training an autoencoder using the local kernelised weight matrix, and 2) fine-tuning the pre-trained auto encoder with the rating matrix, produced by the global convolutional kernel. This repo provide the benmark with the processed data from Movie Recommender Systems.

2. Setup

Download this repository. As the code format is .ipynb, there are no settings but the Jupyter notebook with GPU.

4. Run

1. Use the data in dir processed_data_for_matrix_completion or create a csv file with the same format.
2. Insert the data path in the main code.
3. Run the notebook and see the result.

3. Requirements

• numpy
• scipy
• torch

Matrix factorization

1. Introduction

More about the method can be found here

2. Setup

Download this repository. As the code format is .ipynb, there are no settings but the Jupyter notebook with GPU.

4. Run

1. Use the data in dir processed_data_for_matrix_completion or create a csv file with the same format.
2. Insert the data path in the main code.
3. Run the notebook and see the result.

3. Requirements

• numpy
• scipy
• torch

Proposal method: Heterogeneous Model

Architecture:

Run:

1. Training: run script .heterogenrous-model/train.py
2. Evaluating: run script .heterogenrous-model/eval.py

Requirements:

• torch
• scikit-learn
• transformers

How to sample

Generate the sampled data in either csv or jsonlines format with the sample.ipynb.