Exploring Adolescent Preferences for Battle Royale Games on Mobile Platforms: A Data-Driven Hybrid Framework Integrating KANO, PLS-SEM, and ANN in the Context of Bangladesh

Abstract

Mobile Battle Royale (BR) games such as PUBG Mobile, Free Fire, and Call of Duty Mobile have gained immense popularity among Bangladeshi adolescents. Despite this development, few quantitative studies have addressed the relative importance of different technical and experiential game features in determining player satisfaction and intention to play. To fill this gap, the present work combines the Kano model with Partial Least Squares (PLS) and Artificial Neural Networks (ANN) to develop an entire consciousness of gaming preferences in adolescents. This is a mixed method study in a three-step cycle. First, 515 valid responses were analyzed through the Kano model to classify 15 BR game features into Must-Be, Performance, Attractive, Indifferent, Reverse or Questionable categories. Second, 308 valid cases were used to test a theory-driven structural model via PLS-SEM, examining relationships among system-level factors (device specifications, game interface & controls, internet connectivity & latency, game content & features, game reviews & recommendations), experiential mediators (smooth gameplay performance, battery optimization, graphics quality), satisfaction, and continuance usage intention. Finally, five ANN sub-models were applied to validate and refine the predictive accuracy of the PLS-SEM findings. Kano analysis revealed that Graphics Quality, Voice Chat Quality, Reward System, Social Features and Anti- Cheat Protection are Must-Be features, while Smooth Gameplay Performance and Battery Optimization are Performance features.