Sistem Rekomendasi Film Menggunakan Metode K-NN

Abstract

Recommendation systems are at the heart of cutting through the noise of online data, helping users find content that matches their preferences. In this regard, the K-Nearest Neighbors (KNN) method stands out as a promising approach. KNN, as a similarity-based algorithm, utilizes information from nearest neighbors to make predictions or recommendations. This research explores the implementation of KNN in developing a film recommendation system with a focus on increasing the accuracy and relevance of recommendations. Related references, such as the evaluation of recommendation systems by Herlocker et al. (2004) and the concept of collaborative filtering by Resnick and Varian (1997), are the basis for understanding and improving the potential of KNN. By detailing the exploration of this concept, it is hoped that this research will provide a comprehensive and holistic view in the development of a film recommendation system. From testing using the performance test of the KNN method, namely accuracy, recall, and precision with the best value of 45.4% for accuracy, 45.4% for recall, and 100% precision so that the algorithm can be applied in the film recommendation system.