Anime Recommendation System Using Bert and Cosine Similarity

With the popularity of anime over time, the number of available shows, genres, titles, and thematic elements make it challenging for users to find content that matches their tastes. Because of the many things, users also feel confused in finding which anime suits their tastes because most of the anime in circulation have types that are like one another. Still, other elements differentiate, so classification is needed. This paper presented a recommendation system for anime by implementing BERT, an advanced approach to NLP tasks. The Anime Recommendation System is designed to provide multiple suggestions for anime titles based on their similarities. In this research, the data used are anime titles and genres to compare their similarities with other anime. The anime title and genre data will be featured and extracted using the BERT method and compared with the cosine similarity method.