Prognosticate Trending Days of Youtube Videos Tags Using K-Nearest Neighbor Algorithm

Abstract

YouTube is a video-sharing website where users may publish, watch, share, and comment on videos and other media. The proliferation of technological gadgets, combined with rapid advancements in technology, has resulted in an increase in trending videos on the platform, where videos and content receive hundreds of thousands, if not millions, of views within minutes of being uploaded and continue to trend throughout the day. This study uses the US YouTube Trending dataset, which includes 130591 occurrences and was acquired from the kaggle repository between August 11, 2020 to May 14, 2022. This study used qualitative and quantitative methods to analyze the YouTube videos dataset, and then performed a predictive analysis on the trending video tags, predicting how a particular video on YouTube might trend in the next two to eight days by predicting the trending of such videos for the next two to eight days and showing their accuracy results using the K-nearest neighbor algorithm (KNN). The model that was utilized to perform the prediction analysis has an accuracy of around 98 percent.