An anime recommender system that hits different
View The Website
The Anime 3x3 is a web-based recommender system that we've developed to give people recommendations based on their submitted 3x3. Since the 3x3 is a collage of anime that describes your unique taste, our program looks at your 9 selected anime and trys to offer you similar shows based on your preference.
The algorithms that we have experimented with for the project include: K-Nearest Neighbors (KNN) and Matrix Factorization with Alternating Least Squares (ALS). However, the current system only uses the Sklearn KNN implementation as we've had the most success with it at the moment. The accuracy and efficieny of our current algorithm leaves much to be desired but it is a system that offers good enough results 50% of the time? Future features may include a deep-learning based approach or a more thorough look at content-based and collaborative filtering methods.
/: where you're reading this...hopefullydata/: a partial/modified dataset of the myanimelist kaggle datasetrecsys/: experimental recommender algorithms that we are/have looked intoutils/: scripts for data-preparation/preprocessing of csv
To get a local copy up and running follow these simple steps.
A list of pip packages are listed in requirements.txt. To set up your environment run the following command
pip install -r requirements.txt-
Clone the repo
git clone https://github.com/Pie31415/thirty-three-anime.git
-
Create an
.envfile with aSECRET_KEY -
Run the app
cd anime33 python manage.py runserver
Distributed under the MIT License. See LICENSE for more information.