Personalized Self-Guided Tour Strategy by Integrating Random Forest Preference Model and Attractions Association Model

Abstract

With the rapid development of tourism, personalized self-guided tours have become an important trend in the tourism market. Since traditional personalized recommendation methods often ignore elements such as users' personal preferences and characteristics of travel destinations, in order to improve the reliability of personalized self-guided tour strategy research. The study fuses the random forest algorithm preference model with the frequent pattern growth algorithm association model to provide personalized self-guided tour strategies. The study firstly utilizes the random forest algorithm to predict the attraction preference selection, and then utilizes the frequent pattern growth algorithm to mine the association relationship among the attractions, so as to provide effective data for strategy formulation. The results indicated that the prediction accuracy, recall, and data processing precision of the strategy model for attraction data were 92.07%, 93.07%, and 81.06%, respectively. The coverage of the strategy model was, however, much higher than that of the comparison model for regional attractions, featured attractions, popular attractions, and specialized attractions, at 78.09%, 85.61%, and 63.26%, respectively. This suggests that the model developed in the study can help meet the goal of the trip and greatly increase the accuracy of attractions in the process of developing a personalized self-guided tour strategy. The study intends to increase user happiness and travel suggestion accuracy in order to strongly promote the growth of the tourism industry.