Interest point recommendation method facing location social network

Abstract

The invention discloses an interest point recommendation method facing a location social network. The interest point recommendation method comprises the steps of: firstly, by fusing a time characteristic in a collaborative filtering algorithm based on a user, obtaining an interest point score based on the time characteristic and user collaborative filtering; then, fusing estimation of interest point popularity on the basis of a time factor into a spatial characteristic to obtain an interest point score based on the spatial characteristic and an interest point popularity characteristic; and finally, carrying out linear combination on two scores to obtain a comprehensive recommendation score of the user on each interest point so as to implement interest point recommendation. The interest point recommendation method disclosed by the invention is helpful for improving recommendation accuracy and overcomes the defect that when conventional interest point recommendation is adopted, or a basic collaborative filtering method is applied or the spatial characteristic is introduced into the basic collaborative filtering algorithm, influence of a time sequence and the interest point popularity characteristic on recommendation accuracy is ignored.

Description

technical field [0001] The invention relates to a method for recommending a point of interest for a location social network, and belongs to the field of social interest recommendation. Background technique [0002] In recent years, location based social networks (LBSNs) have developed rapidly, providing multi-dimensional information such as user information, social relations, location coordinates, check-in time, and comment information for recommendation services. In LBSN, users publish their current location information through check-in, and share their comment information and experience feeling on current points of interest (POI, such as tourist attractions, museums, libraries, restaurants, etc.). These location-based social networking sites collect a large amount of user check-in information, and use the user's check-in information to recommend places of interest to users that they have not been to. Point-of-interest recommendation plays a very important role for both us...

AI Technical Summary

Problems solved by technology

[0005] 1) Low recommendation precision and recall

It is difficult to get effective information from check-in data and too sparse data is one of the reasons for low recommendation accuracy

In addition, in-depth analysis of user sign-in data was not possible;

[0006] 2) Dividing time into multiple segments will exacerbate data sparseness

Images