Machine Learning assignment about making an app that recommends anime using KNN principles

So, your a new anime fan, eh? Or you're an oldie that's just returning? Or a recent vet that wanted to explore the deep end? Anyway, this is the app to help you discover new old anime. This app is made based on a Machine Learning algorithm that checks how close the audience is from one anime to another. Its goal is to help you finding new apps that might catch your fancy based on the anime you've watched. Like Bleach? How about Naruto and similiar classic Shonen. More touching cinematographies like Kimi No Nawa? How about Makoto Shinkai's other works, or even the similiarly touching Koe no Katachi. Cool right?

The ARS uses Graph-Based Collaborative Filtering algorithm that measures distance with Cosine Similiarity on the anime's audience and plots the 100 nearest neighboring anime as neigbor nodes. This graph is then stored as an adjacency list model and used to aggregate a list of recommended anime to give to the user.

The data this algortihm is trained on is from 2020, so it doesn't include anime from after that year. This escpecially means it doesn't consider Isekai, the major genre that only became major after the year 2020. It also doesn't include ratings from after that year, and that means no ratings from the lockdown period where people are most likely to watch anime, and it also before the boom of anime becoming mainstream. All this to say, yeah I need a better database because this one is severly outdated.

For the editing software, I personally just used IDLE that comes prepackage with python. That said, IDLE is very primitive and you should probably either use Jupyter Notebook or VSCode. I've never used those for python though so you'll have to learn them through tutorials.

First, before anything else, install the latest version of Python. Then you also need to download the dataset through Kaggle here. Put the dataset into the git folder so that it's positioned in /anime_recommender_system/Dataset/[contents of the dataset].

After all that, run module_install.bat to install relevant packages for this algorithm. Then run main.py. That program will filter the data and clean it before feeding it to the algorithm. After that, run knn.py that would actually train the file.

If you want to get a preview of the data, DO NOT do it through the .csv file. It's way to unreadable for the human eye. Instead, uncomment the lines where it says write_[type]_data. Those lines will call the function to write the data into txt files, which you can then read in Notepad++. No, normal Notepad can't handle those files either.