Sistem Rekomendasi Anime dengan Metode Content Based Filtering

Abstract

Anime is a term for animated films or cartoons produced by the Japanese state. Currently the number of anime in circulation is very large, so anime lovers sometimes struggle to find an anime that suits their tastes. One of the reasons is the limited description and review translated from Japanese into other languages. Making an anime recommendation system with a content based filtering approach that utilizes TF-IDF and cosine similarity. The “genre” feature is used as a recommendation system parameter that will be processed by TF-IDF and cosine similarity. The training data uses data downloaded from Kaggle. Modeling begins by calculating the weight of the genre feature values ??using TF-IDF and looking for similarity values ??using cosine similarity. After that, the process carried out is sorting the similarity values ??on the recommendation system that will display the results of anime recommendations. There is an evaluation of the model, which results in a precision value of 88.1%