This is a content based anime recommendation system builty on myanimelist.net database available in Kaggle.
The trained model is deployed as a streamlit web app:
https://bit.ly/anime_recommend
Developed by: Marcelo Vargas dos Santos
Project content:
- similarity_model.ipynb : In this jupyter notebook I developed a model based on cosine similarity distance of the sinopses vectorized using the ti-idf nlp model.
- index.py : This is the script used by streamlit to deploy the webapp.
- utils.py : Implements utility functions, such as a web scraping to get an anime info from myanimelist.net.