Review Prediction using Large-scale Language Models for Serendipity-Oriented Tourist Spot Recommendation and its Evaluation
Feng Guan,Daisuke Kitayama
18th International Conference on Ubiquitous Information Management and Communication (IMCOM 2024), S4-10, Kuala Lumpur, Malaysia (Online), 2024年1月, 査読有
論文PDF
概要
Existing recommender systems often make recommendations based on the previous actions and preferences of the user, and do not always provide new experiences. Recommendations based on the concept of serendipity can be adapted to such problems. However, in such recommendations, the user may not be able to evaluate the value of the recommended content correctly; this may result in non-acceptance of the recommendation. To increase acceptability, we believe that reviews that users are likely to write themselves (known as predicted reviews) may be more interesting than reviews written by others.