Travel Attractions Recommendation Based on Attentive Group Recommendation Algorithm

In information systems, recommending content to groups of users is essential due to widespread group activities. A key challenge in group recommendation is aggregating member preferences for group decisions. Existing systems often rely on fixed methods for preference aggregation, which can lead to suboptimal outcomes. In this paper, we present an innovative solution named AGREE [1] (short for “Attentive Group REcommEndation”). This approach is built upon attention networks and neural collaborative filtering. AGREE has improved the representation of groups, enabling the learning of interaction patterns between groups and items. We also incorporate user-item interactions, strengthening recommendations for not only groups but also individual users. Experimental results on the data of the real-world show that our model enhances group and user recommendations, particularly benefiting cold-start users without prior interaction history.