Abstract: Aiming at the problem that most of the current locationbased recommendation algorithms did not consider the user’s realtime location and time factors, which made the recommendation lack of realtime, we proposed a recommendation algorithm based on regional active users to solve the cold start problem in traditional locationbased recommendation algorithm. Firstly, the algorithm explored the active user in the area where the user was currently located, and used this as a standard to recommend for the user. Secondly, time factor was introduced into location recommendation to make the recommendation more accurate and realtime. The experimental results show that the number of active user’s friends in the algorithm can improve the realtime performance of the recommendation system, and the fusion of location information in the algorithm can make the recommendation more accurate.

Key words: regional active user, recommendation algorithm, friend recommendation, location recommendation, realtime