https://github.com/pandada8/wooflix
https://gustavonarea.net/blog/posts/korens-svd-python-implementation/
https://github.com/pandada8/wooflix
Last synced: about 2 months ago
JSON representation
https://gustavonarea.net/blog/posts/korens-svd-python-implementation/
- Host: GitHub
- URL: https://github.com/pandada8/wooflix
- Owner: pandada8
- Created: 2016-05-17T18:30:10.000Z (about 9 years ago)
- Default Branch: master
- Last Pushed: 2016-05-17T18:37:09.000Z (about 9 years ago)
- Last Synced: 2025-02-15T18:38:33.444Z (3 months ago)
- Language: Python
- Homepage:
- Size: 208 KB
- Stars: 1
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.html
Awesome Lists containing this project
README
Readme for Wooflix
Readme for Wooflix
Wooflix is a Python-based
implementation of a recommender system for the Netflix Prize, based on Yehuda Koren's
SVD++ model as published on the paper entitled "Factorization Meets the Neighborhood: a Multifaceted Collaborative Filtering
Model".
The HOWTO explains how to install and use Wooflix,
and there's is also a software design document
included.
The /src directory contains the source code for Wooflix,
which is available under the terms of the GNU General Public License v3 or
later.
The software was developed in 11 hours 50 minutes approximately, distributed
as follows (approximately too):
- 6 hours 20 minutes planning the design and writing the design document.
- 5 hours implementing the software in Python.*
- 30 minutes writing the end-users HOWTO.
* It is worth noting that many, many hours were spent waiting for the training
to complete (even with a very small sub-set of the entire data set), but they
were not included in the time spent writing the software.
Another issue that is worth mentioning is that the design document required
the system to train the system with 60 factors and 200 epochs, but these
numbers had to be decreased notably (to 2 and 3, respectively) in order to
make the system run faster (at least in an usable way).