ADAPTIVE PERSONALIZATION OF SOCIAL MEDIA FEED USING POSTCATEGORIZATION

Abstract

In today’s world, social media constitutes a significant part of everyone’s lives. It occupies so much of our time that it even kills our productivity. Social media applications consist of enormous number of posts that can be both informative and entertaining and belong to a wide range of categories. They can be made more user-friendly by the personalization and customization of feeds of users who consume it. Suggestions shown often consist of uninteresting posts, which can make the user experience bad and result in longer scrolling durations. The solution proposed here provides a customized and personalized feed for users based on user interests. Most of the social media applications are proprietary and the algorithms for adaptive personalization are not available publicly. The main objective of this work is to develop a model for deciding how to include or discard a new post from the users' feeds based on user interests collected earlier. The system uses concepts of clustering, informationretrieval and user profiling.