A personalized point of interest (POI) recommendation based on location social network has important practical significance for users and businesses. To address the problem of lower accuracy in the personalized recommendation algorithm, this paper applies composite time and location information to the user-based collaborative filtering algorithm and proposes a fusion multidimensional information personalized POI recommendation algorithm-RecUTG. This algorithm mines the interest preferences of the user in a different context dynamically, calculates the interest weights from the check-in time and location perspective, finally learns collaborative filtering ideas to predict the rating of a user, and then makes recommendations according to the interests of the user. Experimental results show that the POI recommendation algorithm has relatively high recommendation accuracy and achieves customer satisfaction compared with existing algorithms.
展开▼
