Penerapan Fuzzy Logic dalam Sistem Rekomendasi Film: Studi Kasus Justice League (2017)

Enhancing the user experience on social networks and e-commerce sites is often supported by accurate recommendation systems. This recommendation task is challenging as it requires a high degree of accuracy. This paper discusses the implementation of fuzzy logic in a movie recommendation system. Two similarity measurement methods are used: the first by considering the choices of similar users, and the second by matching genres of movies that have been rated by the user. To calculate similarity, four different techniques are used, namely Euclidean Distance, Manhattan Distance, Pearson Coefficient, and Cosine Similarity. Fuzzy Inference System (FIS) and Adaptive Neuro-Fuzzy Inference System (ANFIS) are applied in decision-making. Experiments are conducted on the Movie Lens dataset, and a comparative performance analysis is reported. Experimental results show that ANFIS outperforms FIS in most cases when the Pearson correlation metric is used for similarity calculation.