JIITA, vol.8 no.4 p.1001-1006 2024, DOI: 10.22664/ISITA.2024.8.4.1001
Hyeonji Kim, Yoosoo Oh
Daegu University
Abstract: The title of a YouTube video is the first thing people see before watching the content. The number of views of a YouTube video differs depending on the level of interest in the title. In this paper, we use a machine learning regression algorithm to analyze the number of views according to the title of Korean YouTube videos. This paper learns YouTube title as the feature value and the number of views as the target value. This paper designs a learning model using the KernelRidge regression algorithm, which achieved the highest performance as a result of analyzing six machine learning regression algorithms (LinearRegression, KNN_Regression, SVR, KernelRidge, DecisionTree Regressor, gradient boosting regressor). The proposed system predicts the number of views through a learned model when a user enters a YouTube title.
Keywords: YouTube, Machine learning, Regression, KernelRidge
Fullpaper: