Recommending Tourist Destinations based on the Distribution of Attributes in Travel Planning via Chatbot
Junsaku Takada, Daisuke Kitayama
The 21st International Symposium on Advanced Technology , ICTL4, オンライン, 2022年11月, 査読無
論文PDF
概要
Planning group travel or tourism events can be challenging for participants with ambiguous wishes or who may have difficulty clearly expressing their opinions. In this study, we developed a chatbot designed to support travel planning by collecting candidate genres and destinations from users at arbitrary times and submitting the candidates to a vote among users. The output is a set of candidates based on the users’ inputs. The candidates to be added are determined by estimating the attributes of the input on which users disagree. Because the voting results reflect the implicit wishes of each user, the proposed method captures the opinions of users who are less likely to express their preferences explicitly. By repeating this process while discussing their plans, users can select tourist destinations to visit. The results of a single test run of our proposed approach show that the number of genres recommended fell with the number of genres entered.