https://github.com/sof3/stat4609-project
STAT4609 project, in collaboration with @jevrii and @kellycyy. Git commit authors do not necessarily represent the actual author.
https://github.com/sof3/stat4609-project
collaborative-filtering knn ncf netflix-prize recommendation-system svd
Last synced: 6 months ago
JSON representation
STAT4609 project, in collaboration with @jevrii and @kellycyy. Git commit authors do not necessarily represent the actual author.
- Host: GitHub
- URL: https://github.com/sof3/stat4609-project
- Owner: SOF3
- Created: 2021-03-19T02:36:33.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2021-05-05T06:24:32.000Z (over 4 years ago)
- Last Synced: 2025-03-20T09:51:19.603Z (9 months ago)
- Topics: collaborative-filtering, knn, ncf, netflix-prize, recommendation-system, svd
- Language: Jupyter Notebook
- Homepage:
- Size: 10.1 MB
- Stars: 2
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# stat4609-project
This is the course project conducted by @jevrii, @kellycyy and me.
The project studies the [Netflix Prize dataset](https://www.kaggle.com/netflix-inc/netflix-prize-data),
developing three models (kNN, SVD and NCF) to predict ratings.
The project report can be found at [report.pdf](./report.pdf).
Code implementations are found in the following files:
- kNN: `knn_full_submit.ipynb`
- SVD: `svd_full_submit.ipynb`
- NCF: `ncf_with_full_dataset.ipynb`
- GMF: `gmf_with_full_dataset.ipynb`
Since this project is only of technical interest of demonstrating different models,
we did not try to push the test RMSE to its best.
The best result we obtained was only around $0.87$.