Enhancing the Quality of Service in Video Game Live Streaming Using Big Data Analytics with DNN Classification and BERT-Based Sentiment Analysis
Sasikumar Perumal, Kalaivani Kathirvelu · Engineering Technology & Applied Science Research · 2025
As live streaming video gaming platforms become more and more popular, Quality of Service (QoS) must be improved to increase viewer engagement and satisfaction. Many factors affect the user experience, including interactivity, streamer behavior, video quality, and active community engagement. Real-time QoS optimization is still difficult considering changing network conditions, audience participation, and content quality. This study explores the behavior and expectations of participants in live streaming video games using big data analytics. Driven by network stability, stream quality, and audience interaction parameters, a Deep Neural Network (DNN) can predict viewer satisfaction and turnover rates. BERT-based sentiment analysis is used to extract trends of audience engagement and general sentiment from chat interactions and comments. The proposed framework can extract important insights that can guide modifications to improve QoS. Experimental evaluations show that the proposed DNN using BERT improves the accuracy of satisfaction prediction using datasets from Twitch, Facebook Gaming, and YouTube Gaming. In response to real-time QoS demands, the model can dynamically alter bitrate and resolution to optimize streaming performance. The results show that the proposed model can provide better viewer retention, reduce buffering problems, and increase engagement to help content creators and platform developers.