Beyond Accuracy: How Algorithm Characteristics and Digital Generational Differences Affect User Interaction and Acceptance in Fashion Recommendation Systems

Algorithm recommendation systems have become core technologies for fashion e-commerce platforms to enhance user experience. However, significant differences exist in interactive behaviors across digital generations. Based on the Technology Acceptance Model (TAM) and Stimulus-Organism-Response (SOR) theory, this study surveyed 560 Chinese fashion e-commerce users and employed Structural Equation Modeling (SEM) to explore how algorithmic recommendation characteristics influence interaction behaviors across digital generations. Results reveal that algorithm accuracy has dual effects: positively associated with perceived usefulness and trust, yet negatively associated with perceived style diversity. Digital generations significantly moderate these effects. Digital Natives rely more on accuracy for trust-building, while Digital Immigrants are more sensitive to transparency and style diversity. This study innovatively introduces “perceived style diversity,” revealing cognitive mechanisms underlying generational differences in user-algorithm interaction. It enriches human-computer interaction theory in algorithmic systems and provides guidance for designing adaptive interfaces and differentiated interaction experiences across generations.