Please download dataset and put into folder named 'data' in the same directory to run model: https://www.kaggle.com/datasets/marlesson/myanimelist-dataset-animes-profiles-reviews
This project is a machine learning recommender system built to recommend Japanese Animated Shows based on your current favorites ! Takes in 130k reviews from MyAnimeList, creating utility matrix and using K-nearest neighbours to train the model.
It recommends you Animes to watch based on public user reviews and data.
- First rerun model
- Select favorite show(s) through lookup in database
- Enter selection and watch ML magic !
.
First run may take a couple of hours - TO MAKE FASTER - decrease size of data: remove_n = int(reviews.shape[0] * 90 / 100) -> remove_n = int(reviews.shape[0] * 99 / 100)