Retweet Reputation: A Bias-Free Evaluation Method for Tweeted Contents

Abstract

The widespread of word of mouth using retweets on Twitter has enabled us to estimate trends in the real world. Previous research methods estimate the value of a tweeted content by calculating the number of subscribers who receive the tweet. However, we should consider the numbers of followers for both the tweeter and retweeter(s) as a greater number of followers may result in more retweets, which we call “bias.” In this paper, we propose a bias-free evaluation method for tweeted contents. Experiments show that our method is successful at evaluating tweets without biases.