Social Media Profile Tagging – A novel machine learning approach for Twitter

Abstract

The project utilizes the sheer number of Twitter profiles and classifies them based on profuse facets. Dominant aspects have been extracted from the users of Twitter that are discernible on a profile such as the number of followers, username, retweets count, and likes count, etc. The proposed system categorizes Twitter profiles into six categories: Political, Actor, Sports, Singer, Educational, and Content Creator. A machine learning approach has been used in the proposed model that works on these abundant features. Several heterogeneous models such as Bayesian networks, SVC, Random Forest, and CNN are used in the study to achieve desirable results. The preferred system is applicable to all Twitter profiles and helps label them through profile URLs. It also helps recommender systems and Twitter to analyze the profiles based on broader categories