Ani-4-Me offers a simple solution for finding new shows to watched based on past shows that you and you friends have watched and overall ratings.
Anime is hand-drawn and computer-generated animation originating from Japan. Ranging from comedy to psychological thrillers, this medium comes in many genres and is consumed by more than a third of the world’s population. However, once people finish watching a show, they often spend a long time de- termining what to watch next. On top of that, one’s anime preference varies from person to person and can be influenced by various factors such as genre, episode number, popularity, and what their friends are watching. While there have been platforms like MyAnimeList, a social networking site that provides its users with an organization system to rate anime and find users who have similar tastes, the lengthy and tedious process of finding a new show, with no guarantee of liking it, is dreaded by anime-watchers.
This project comes with the goal of giving providing relevant recommen- dations for new anime to watch. Rather than spending time surfing the web and reading reviews with no guarantee of finding an enjoyable anime to watch next, I have created a recommendation system for anime. There are 2 main methods for implementing recommender systems, which are collaborative filtering and content-based methods. Collaborative filtering involves looking at past user-item interactions and suggesting new items based on their proximity, while content-based methods use information about the user and recommend items based on how well the items match with the user’s given information. A combination of the two algorithms is used to provide an accurate recommendation for users. This algorithm reccomends anime based on the previous anime that you have watched, the ratings that you have given them, and your preferences for things such as genres, the number of episodes, and the start and end dates of the airing.
The dataset our group used is the Anime Dataset with Reviews - MyAnimeList from Kaggle. This project mainly uses trees, dictionaries, and sets to represent our problem domain. Graphs are an important part of our domain as they helped represent the connections between users and animes through review edges and between users as friend edges. Graphs are also important to a part of our recom- mendation algorithm since part of the algorithm is based on doing a depth-based DFS to find recommended anime.