Site recommendation method based on time and geographical location information integrally

Abstract

The invention discloses a site recommendation method based on time and geographical location information integrally. The site recommendation method based on time and geographical location information integrally comprises the following steps: 1, acquiring site attendance data of a user; 2, carrying out multi-center clustering on historical attendance sites of the user to obtain an active zone of the user, and calculating the probability that the user arrives at an unattended site on the basis of the distance from the center of the active zone of the user to the unattended site of the user; 3, calculating similarity of various time periods, filling up the attendance times of the user at the unattended site according to the similarity of the time periods by using existing attendance data, and calculating site similarity under time perception according to filled-up attendance data; 4, predicting the attendance times of the user at the unattended site on the basis of the probability that the user arrives at the unattended site and the site similarity under the time perception; and 5, recommending sites with top N predicted values for every use according to the predicted values in the step 4 so as to generate a final recommendation result. By the site recommendation method, the accuracy and the recall rate site recommendation can be improved.

Description

technical field [0001] The invention relates to a method for location recommendation integrating time and location information, belonging to the technical field of user recommendation systems. Background technique [0002] In recent years, with the rise of emerging e-commerce services, the explosive growth of Internet information, the variety and complexity of information, and users have more choices. However, due to information overload, the user's satisfaction is reduced, and it becomes more and more difficult for users to select the items they are interested in from the numerous and complicated items (products and services). Recommender systems are an effective tool to address this information overload problem. On the one hand, the recommendation system can help users discover valuable information for themselves, and on the other hand, the information can be displayed in front of users who are interested in it, so that producers and consumers can achieve a win-win effect...

AI Technical Summary

Problems solved by technology

[0006] (1) In the location recommendation, the collaborative filtering algorithm only uses the number of check-ins of the user in the location to make recommendations, but due to the sparsity of user check-in data, the recommendation accuracy is poor;

[0007] (2) In the existing theory, when calculating the probability of a user arriving at an unchecked location based on geographic location information, all the information of the user’s checked-in location is used indiscriminately, which will increase the complexity of the calculation on the one hand, and on the other hand, if Users may occasionally visit some remote locations and have low-frequency check-ins, but these low-frequency locations are not frequently visited by users. Considering these locations indiscriminately will affect the results of probability calculations and reduce the accuracy of recommendation results.

Images