Enhancing Cloud Gaming Experience: Video Quality Prediction with User Activity

Abstract

With the rapid development of cloud gaming technology over the past twenty years, cloud gaming has begun to challenge traditional gaming modes. Facing an increasing number of cloud gaming platforms, reducing service costs and improving service quality have become key focus areas. This study proposes an innovative cloud game video quality prediction model. It aims to predict the future trend of video quality by analyzing simulated cloud game players’ interaction data and cloud game streaming video quality parameters. In our research, by collecting and analyzing multiple sets of relevant data, we established the Linear Time Series Analysis Model and the Recurrent Neural Networks Based Model, and compared them to derive a new prediction model for cloud game streaming video quality. This model not only predicts the future trend of video quality to optimize user experience and service cost, but also derives multiple targeted models to adapt to the specific needs of different game genres through users’ behaviors in different games. Through simulation experiments, we verified that these models are effective in various game scenarios and can accurately predict changes in cloud game streaming video quality. The research results demonstrate the applicability and accuracy of our model for cloud game video stream quality prediction, providing cloud game service providers with strategies to optimize video quality based on user behavior dynamically. In addition, this study explores the potential applications of the prediction model in streaming media and cloud gaming services, providing theoretical support for the future direction and improvement of cloud gaming services. Finally, we anticipate that further refinement of user behavior analysis could enhance the performance of the prediction model and the cloud gaming experience. The outcomes of this research offer strong support for improving the quality and efficiency of cloud gaming services, and have practical application value and theoretical significance. Key words: Cloud Gaming, Video Quality Prediction, Video Multimethod Assessment Fusion, User Activity Index, Time Series Models, Nonlinear AutoRegressive Model with eXogenous Inputs, Recurrent Neural Networks, Quality of Service, Quality of Experience