Dynamic Predictive Modeling Approach of User Behavior in Virtual Reality based Application
Ahmet Köse, Aleksei Tepljakov, Eduard Petlenkov · 2019
Virtual Reality (VR) is considered to be a powerful modern medium for immersive data visualization and exploration. However, few studies have proposed solutions to complement data visualization in immersive environment considering the user's behavior. This paper addresses dynamic modeling of user behavior approach in an interactive VR based application. In this application, real-time data communication is employed to track accurate location and orientation of head mounted display device worn by the user. In our experiment, we use example of collected data and provide a methodology to predict next movements of the user by using nonlinear autoregressive (NAR) and location in the application by the nonlinear autoregressive neural network with exogenous inputs (NARX). Results suggest both neural networks are suitable for performing prediction which can be used to achieve an improved feeling of presence while reducing required high computational power. Data analysis part of the research is also linked to human behaviors to improve studies which are usually performed by traditional survey techniques.