Detecting Outlier Behavior of Game Player Players Using Multimodal Physiology Data

Abstract

This paper describes an outlier detection system based on a multimodal physiology data clustering algorithm in a PC gaming environment. The goal of this system is to provide information on a game player's abnormal behavior with a bio-signal analysis. Using this information, the game platform can easily identify players with abnormal behavior in specific events. To do this, we propose a mouse device that measures the wearer's skin conductivity, temperature, and motion. We also suggest a Dynamic Time Warping (DTW) based clustering algorithm. The developed system examines the biometric information of 50 players in a bullet dodge game. This paper confirms that a mouse coupled with a physiology multimodal system is useful for detecting outlier behavior of game players in a non-intrusive way.