Generate Personalized Explanations for Recommendation based on Keywords
Wei Guo Song, Chenglong Wang, Keqing Ning · 2021
Explainable recommendation refers to providing users with recommended products and explaining to users the reasons for recommending the products. Recommendation explanations can greatly increase users’ trust and satisfaction with the recommender system, and to a certain extent can assist users make decisions efficiently. The current recommendation explanation is mainly templated sentences although this method is simple and easy to understand, it is relatively rigid, lacks flexibility, insufficient service, and requires a lot of manpower and material resources. Inspired by the above questions, by mining user comment information, we propose a method to generate multiple recommendations based on keywords. First, the keywords in the comment information are extracted through STF-IDF, and then the recommendation explanation is generated through the classic network GRU generated by natural language. Experiments show that our proposed method not only has better recommendation accuracy but is also has a higher quality of recommended interpretation compared to classic methods