Advanced Algorithmic Techniques for Topic Prediction and Recommendation—An Analysis

With the rapid growth in technology and the need for people to stay connected, social networking has become the most essential part of day-to-day life. As the demand went high, a lot of social networking applications were made available to the users. Twitter is one such famous platform where users get to post short and crisp texts called tweets. Being restricted with a limit of 140 characters per tweet is one of the effective ways to diffuse information. Twitter has a unique ability to organize and encourage large-scale debates about ‘trends.‘ Users can stay updated with the latest trends around the world by using twitter’s large-scale platform of posting tweets, which are updated in real-time. Twitter's uniqueness from other social media platforms comes from its ability to list out hot trending topics being discussed by a large number of people. Predicting the topics which are soon be trending is a significant way to help users who want to grow their following by communicating their message to others. In this paper, a recommendation based on pre-trending subjects based on individual user interests, as well as prediction algorithms for identifying tweets that will shortly trend has been detailed. Techniques like Hashtag, Word Ranking, Authority, Background Tweet Detection and Short-Term Fluctuation Modeling based Approaches are analyzed to show the performance of predictions.