Identifying factors influencing satisfaction in self-drive tour using random forest and importance-performance analysis

This study examines the integrated experience of self-drive tour, with a focus on both driving and sightseeing activities. It aims to systematically identify the key determinants of tourists’ overall satisfaction and to inform service resource optimization strategies. Data were collected through an online questionnaire survey, and a random forest model was applied to analyze multiple scenarios involving road travel and scenic attraction experiences, thereby revealing the nonlinear effects of influencing factors. The results indicate that scenic area visitor flow, vehicle automation level, age, parking fee, and self-drive frequency are significant predictors of overall satisfaction. By incorporating Importance–Performance Analysis (IPA), the study further clarifies priorities for resource allocation. Vehicle automation level, scenic area visitor flow, and age-related services fall within the ‘Concentrate here’ area and require prioritized resource allocation. Parking fee and self-drive frequency belong to the ‘Keep up the good work’ area, where existing services should be maintained and optimized. In contrast, one-way travel distance, vehicle type, and travel group size are classified into the ‘Possible overkill’ area, indicating potential for differentiated service development. This study provides empirical evidence to support managerial decision-making under conditions of limited resources.