Measuring the Information-Foraging Behaviors of Social Bots Through Word Usage

Abstract

Automated social bots are reported to account for a large sum of activity on social media sites such as Twitter. In this short paper, we study the information-foraging behaviors of social media users including bots. We present here a preliminary investigation which compares the behaviors of a set of suspected bots with non-automated accounts. To do so, we measure the distance between word distributions on a daily basis. We posit that this methodology provides a quantitative measure of behavior, which allows for more rigorous descriptions of bot behaviors that move beyond the assumption of bots as a monolithic category.