A smart rural tourism resources recommendation based on audience preference

How to provide users with more accurate smart rural tourism recommendation services has become a hot research topic at present. To address the short-term audience preference issue caused by data scarcity, firstly, graph convolutional networks (GCN) are applied to recommend smart rural tourism resources. For long-term tourism audiences with sufficient data, use long short-term memory (LSTM) to construct a recommendation model based on users' long-term dynamic preferences. The results showed that in the case of data scarcity, the recall and accuracy of the GCN recommendation method increased by 17.9% and 11.8%, respectively. In long-term rural tourism applications, the hits ratio (HR)@10 and HR@20 of the dynamic preference recommendation model were as high as 42% and 50%, respectively. The results indicate that the proposed method provides more reliable technical support for intelligent rural tourism recommendation and can more effectively discover audience preferences.