Point-of-interest recommendation method that integrates text and geographic information in local co-arrangement

Abstract

The invention discloses a point-of-interest recommendation method fusing a text and geographic information in local synergistic arrangement. The method comprises the following steps of S1, inputting a training set and randomly generating anchor points; S2, calculating the similarity between a user POI (point-of-interest) and each anchor point according to an implicit eigenvector, a spatial characteristic and a topic characteristic; S3, obtaining the similarity according to the previous step, finding a (user-POI) pair as a member in a sub-matrix, and randomly initializing a user eigenvector and a POI eigenvector for each sub-matrix, wherein each anchor point represents one sub-matrix, and the similarity between the (user-POI) pair and the anchor point is greater than a preset threshold; S4, acquiring a user and a position point; S5, calculating a gradient and updating the user eigenvector and the POI eigenvector; S6, judging whether a iteration stop condition is met or not, and if yes, skipping to the step S7, or otherwise, skipping to the step S4; S7, estimating preference values of the user for the POI according to the user eigenvector and the POI eigenvector; and S8, sorting the preference values and generating a recommendation list for the user.

Description

technical field [0001] The present invention relates to the field of information push, and more specifically, relates to a point-of-interest recommendation method that integrates text and geographic information into a local collaborative arrangement. Background technique [0002] In the social network based on location services, there are a large number of points of interest (point of interest, POI, or location points), such as restaurants, hotels, scenic spots, etc., and users often need to face tens of thousands of location points to make a decision. choose. How to help users filter out useless information, find out the points of interest that users may care about or like the most, and recommend them to users, this is the work that the point-of-interest recommendation system must complete. [0003] In terms of point of interest recommendation, the main problem is how to solve the problem of poor recommendation effect caused by sparse data. And in the face of multi-source...

AI Technical Summary

Problems solved by technology

This method only considers user check-in data, and geographic and text data are not well integrated into the model

Images