Classifying Twitter Users Based on User Profile and Followers Distribution

Abstract

We propose methods to classify Twitter users into open accounts and closed accounts. Open accounts (shop accounts, etc.) are the accounts who publish information to general public and their intentions is to promotion products, services or themselves. On the other hand, closed accounts tweet information on their daily lives or use Twitter as a communication tool with their friends. To distinguish these two different kinds of Twitter users can help us to search for local and daily information on Twitter. We classify Twitter accounts based on user profiles and followers distributions. The features of profile of open accounts include clue keywords, telephone number, detailed address, and so on. Follower distribution is another notable feature: most open accounts have followers from variety community. The experimental results validate our methods.