Multiple Time-Periods POI Recommendation Model Based on Distributed ELM

To solve the problems of low efficiency of algorithm execution and not fully considering the POI needs of users in different time-periods in point-of-interest (POI) recommendation. In this paper, propose a Multiple time-periods POI recommendation model MP-DELM based on distributed extreme learning machine (DELM). According to the user's different periods and different types of POI access requirements, the data set was processed in different periods; extract POI preference features: distance preference feature, personalized preference feature, popularity preference feature, and friend preference feature, in which for the friend preference feature, social friend feature and the trajectory feature are integrated for the first time, and a new trajectory friend similarity calculation method is proposed; in order to improve the efficiency of POI recommendation, improve DELM classifier. And the online sequential extreme learning machine (OS-ELM) is used to update training data in real time. Finally, we prove that the MP-DELM recommendation model can effectively improve the accuracy and efficiency of POI recommendation in the real datasets.