Analisa Sistem Rekomendasi Konten Youtube Berdasarkan Durasi Menonton Menggunakan Content-Based Filtering

Abstract

In today's era, the internet is a facility in social life that causes phobias, complete and necessary information is difficult to obtain again. However, how YouTube provides consistently using an algorithm designed for YouTube content recommendations which is an online video media can witness important moments instantly to individuals who are not on television media so that all users can get useful information and entertainment from website media. For some reason Youtube is used as social media with the highest user level from Instagram. Therefore, we make an experiment to categorize the right content to be a crucial factor in producing accurate and meaningful recommendations. In a system analysis, it recommends content on Youtube based on individual categories using the basic concept of the content based filtering algorithm and how it is implemented in the context of YouTube. The model training is carried out using the cosine similarity method which aims to compare the similarity between the contents of these representations. Evaluation of the model can provide insight into how effective the algorithm is in producing relevant recommendations. The steps in the recommendation system analysis are literature study, data collection, model training, and model evaluation by increasing understanding of content-based filtering algorithms.