User Perception and Evaluation of a Deep Learning Framework for Audience Engagement Analysis in Mass Events

As video volume on the web grows exponentially over time, video streaming platforms have been enhanced with Artificial Intelligence (AI) to analyze their content. This paper proposes a novel evaluation methodology for video events streaming platforms that useAI for content analysis and helps themmeasure user experience, customer satisfaction, and AI acceptance. Models like System Usability Scale (SUS), Technology Acceptance Model (TAM), European Customer Satisfaction Index (ECSI), and Net Promoter Score (NPS) have been fused, creating a novel methodology for the evaluation. To this end, correlations between items, model scores, and statistic metrics were utilized. Experimental results in a real AI-enabled video streaming platform verified the potential of this evaluation methodology, with insightful conclusions drawn from it. The study helps in similar evaluation tasks and provides crucial information to software and system designers who want to know where to emphasize and how to evaluate similar systems. Results provide rich details on the platform's user experience, demonstrating how important it is to enhance online video streaming platforms with AI analysis features.