Computing and Informatics

With the continuous development of smart mobile devices, user check-in behaviors data has experienced explosive growth. To solve the problem that existing point-of-interest (POI) recommendation models fail to fully consider the complex and dynamic changes in user preference behaviors, a POI recommendation model based on a Bi-directional Long Short-Term Memory recurrent neural network (Bi-LSTM) and an attention mechanism is proposed. By integrating the time sequence and spatial location information of user check-in behaviors into Bi-LSTM, the user's spatio-temporal feature sequential dependence is established. The attention mechanism is utilized to model historical check-in locations, derive a weight sequence, and recommend the next POI to the user based on the probability distribution of predicted POIs. Experimental results on two public datasets demonstrate that the model outperforms previous advanced POI recommendation models in multiple evaluation metrics and has better recommendation performance.