Overview of Present Methods to Predict Social Media Metrics of a Post

Social Media is an ever-growing tool which will only continue to grow as time goes by. As the commercial applications of social media increase exponentially, interest in using AI techniques to predict the performance of posts has also increased. In this paper we have discussed four major approaches and methodologies to analyse the information present in social media posts and on the basis of the historical data obtained, predict the popularity of any future post in terms of number of likes, comments and shares. These approaches include Deep Neural Network (DNN), Multimodal approach by using the multimodal information embedded in a social media post on a given social media platform, Social Network Services (SNS) for the collection and inference of user-related data and Social Forecast method for the accurate forecast of popularity of content put up on a social media platform. All these methodologies are performed on a dataset in the form of algorithms to obtain results of predictions. These methods take into consideration textual and multimedia content such as images and videos and on the basis of data acquired from these, eventually draw results. Our hope is that this paper can provide an insight into the current state of ML approaches towards social media metrics prediction.