Comparative Analysis of Recommendation System for Social Media Analysis

Abstract

Since the growing population uses social networks in their everyday lives, so its generated data is analyzed and visualized in various fields. The way of production, transfer, and consumption of the information, is represented by online social media. Elements like tweets, posts, and comments, set up a connection between the maker and the user of the generated data. Social networks like Facebook generate 4 petabytes of data per day, 511,200 tweets per minute on Twitter, 240MB of data usage per hour on YouTube.

Predictive decisions are taken using social data in many applications such as e-commerce, business, travel, news, micro-blogging, etc. Thus, the social media website facilitates linking people, having similarities, exchanging ideas, and forming groups, and encouraging many social and commercial activities. Due to the information overload, finding the correct information gets demanding. The recommender system is a type of refining system that filters primary information, considering the user choices or feedbacks. The main purpose of the analysis is to give an outline of the recommender system, its types, and the related algorithm. We present, in this analysis, a review of some papers published between 2013 and 2019 in the social media field. Finally, we have mentioned their key aspects with employed techniques. Due to the explosion of social media data, we focus on the recommender system study. We will conclude by pointing out a set of machine learning algorithms that will be used for the contributions towards future work.