A Two-Tiered Recommender System for Tourism Product Recommendations

Existing travel recommender systems have difficulty in automating word-of-mouth communication and tend to only mimic the role of traditional travel agents. This study proposes a novel travel recommender system based on collaboration filtering and approximate constraint satisfaction that can automate word-of-mouth communication and provide personalized travel services. The proposed travel recommender system models a tourist's personal needs as a constraint satisfaction problem and helps the user build a personalized travel plan. However, because the existing constraint satisfaction method is often too rigid, this study adopts an approximate constraint satisfaction method by incorporating indifference intervals into constraints. We implement a prototype system and verify the effectiveness and usability of the system. The experiment results show that it is a promising system for the automation of word-of-mouth communication on the destination and user-defined travel planning service.