Transcending Value Judgments: Differentiated Travel Demand Strategies Driven by Deep Neural Networks From the Perspective of Consumer Perceived Usefulness

Abstract

In the context of rapid digital tourism development and platform-based competition, online travel platforms (OTPs) increasingly exhibit coexisting functional convergence and intensified rivalry. This makes it critical for platforms to identify sources of user-perceived differentiation embedded in user-generated content (UGC) and translate them into actionable competitive diagnostics. Focusing on mobile OTP applications, this study analyzes 95,960 online consumer reviews (OCR) collected from Apple App Store and develops a human-in-the-loop, multi-algorithm text-mining framework. By integrating information-entropy filtering, semantic representation and keyword extraction, and deep-learning-based sentiment computation, we identify and quantify consumers’ feature salience and sentiment feedback across five decision-relevant dimensions: price, service, travel activities, accommodation & transportation, and software interaction. Building on these outputs, we bridge social proof theory and the resource-based view by proposing a value–scarcity dual-criterion competitiveness matrix that converts consumer perceptions into diagnostic cues for differentiation-oriented resource configuration. This study contributes a replicable pathway from large-scale UGC to quantified strategic insights, extends social proof from individual decision-making to platform competition and strategic diagnosis, and offers data-driven implications for improving platform mechanisms and experience design under homogenized competition.