G-DAN-SNR: An improved Personalized POI recommendation algorithm based on gating and attention mechanism

In the field of POI recommendation, some scholars proposed the DAN-SNR model, which constructs the deep learning network by self-attention mechanism and designs dual channels to capture users' long and short interest preferences and social influence. Based on the improvement of the DAN-SNR model, the gating mechanism is introduced in this paper based on the self-attentive mechanism before its long- and short-term interest preference features are fused, and the optimal features are filtered by weighting the feature parameters.