Abstract: Nowadays, among the many social media platforms, the number of individuals earning a livelihood through the YouTube social media platform is rapidly increasing. In this situation, YouTube also has many channels and categories, and what is trending can vary depending on the country and year. In making a living with this platform, choosing content that is likely to be trending can help you become a successful YouTuber in a short time. It is proposed with the aim of having a significant impact on content analysts. This system is implemented by utilizing a blend of trend analysis and sentiment analysis to select the appropriate career path, with more accurate and better results. The trending analysis involves developing a new trending algorithm inspired by YouTube’s trending system and analyzing the content of trending YouTube videos. Sentiment analysis is implemented using a combination of TF-IDF and Multinomial Naive Bayes models. In terms of performance, the accuracy of both analyses is good, and the results are accurate. For the trending analysis, linear regression and logistic regression were used as two approaches to calculate performance, and an accuracy value of up to 91% was achieved. The performance classification result for sentiment analysis was calculated at up to 89.54%, and it was compared with the ground truth to demonstrate its accuracy. This system uses India YouTube trending videos from Kaggle and will be very helpful and beneficial for YouTube career seekers, content analysts, and professional YouTubers.

Keywords: YouTube Trending Analysis, Sentiment Analysis, Statistical Count, Linear Regression, Logistics Regression, Multinomial Naïve Baye, TF-IDF, Pearson's correlation Analysis

Pages: 271-286