Trending or not? Predictive analysis for youtube videos

The internet has brought about significant transformations in communication and human behaviour. It has revolutionised how people connect, express themselves, and socialise. The emergence of social media platforms has turned the virtual world into a tangible reality, enabling global connectivity without barriers. Initially designed to stay in touch with acquaintances and share thoughts, social media platforms have evolved to offer a diverse range of services. These platforms have become thriving marketplaces, influencing consumer behaviour in various ways. Platforms like YouTube have witnessed notable changes in the number and nature of advertisements accompanying videos. Due to the revenue-sharing model based on advertisements, YouTube had to increase its video view count. Additionally, YouTube has introduced video categorisation, with one such category being "trending videos." This proposal utilises existing machine learning techniques like support vector, logistic regression and decision tree to predict whether a video will be categorised as trending, employing supervised machine learning methods. The results are then compared based on their accuracy and precision, providing insights into the effectiveness of the techniques.