Predicting digital heritage engagement through cultural identity: a hybrid SEM-ANN study of Digital Dunhuang

Digital cultural heritage (DCH) platforms support heritage interpretation and public engagement, but the psychological processes associated with users’ continuance and recommendation intentions remain unclear. Drawing on the cognition-emotion-behavior (CEB) framework, this study used structural equation modeling (SEM) and artificial neural networks (ANN) to analyze survey data from 504 respondents who had used Digital Dunhuang or comparable DCH platforms. SEM showed that experiential dimensions were positively associated with emotional experience, satisfaction, and cultural identity; these variables were positively associated with both intentions. Cultural identity was the strongest predictor of both outcomes among the three psychological variables. Mediation models indicated partial mediation through emotional experience, satisfaction, and cultural identity. ANN sensitivity analysis broadly supported the SEM rankings and quantified relative predictor importance under nonlinear modeling. The findings identify cultural identity as a central correlate of intended digital heritage engagement and inform DCH platform design and management.