Method and system for recommending Geo-tagged items

• US 8,793,248 B2
• Filed: 04/23/2012
• Issued: 07/29/2014
• Est. Priority Date: 04/26/2011
• Status: Expired due to Fees

• Current Assignee: Yoochoose GmbH (eZ Systems AS)
• Original Assignee: Yoochoose GmbH (eZ Systems AS)
• Inventors: Friedmann, Michael, Ben-Shimon, David, Rokach, Lior
• Primary Examiner(s): CHANNAVAJJALA, SRIRAMA T
• Application Number:
  US13/453,040
• Publication Number:
  US 20120303626A1
• Time in Patent Office: 827 Days
• Field of Search: 709/203, 709/206, 709215-220, 707/688, 707705-709, 707720-728, 707748-754, 707731-732, 707/758, 705/7.34, 705/14.6, 705/26.1, 705/26.64
• US Class Current: 707/732
• CPC Class Codes: G06F 15/16 Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs coordinating program control therefor G06F9/52; in regulating and control system G05B
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06Q 30/00 Commerce
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A recommender system for recommending items to a user based on geo-Tagged information related to him, in which items associated with a GeoTag are stored in a database. Feedback regarding the various items is obtained from the user and the provided rating of items is propagated to closely located items based on their associated GeoTags. A user-to-user similarity matrix is calculated and a predicted score is assigned for each user and item, using a recommendation server. All the items in the catalog of items are sorted according to their predicted scores as calculated for the user, and all items that have been already rated by the user are filtered out. Then, items from the catalog of items are presented to the user, according to their scores.

Claims

1. A recommender system for automatically recommending items to at least one electronic device of one or more users, based on geo-tagged information related to said user, comprising:

a database for storing data in a non-volatile storage, each of which being associated with a GeoTag;

a recommendation server, in communication with said database and being connected to said electronic devices via a client application and over a data network, configured to;

obtain, over said data network, rating values from the user regarding geo-tagged items stored in said database in a form of a user-item matrix;

form weighted rating values for said items being closer than a predetermined geographical distance, wherein said weighted rating values are weighted according to;

a geographical distance between a pair of items;

the items previously rated by said user;

a level of similarity between the user and other users; and

preferences of similar other users to said user;

propagate a user-item matrix with the weighted rating values;

construct a user-to-user similarity matrix by calculating the similarity for each pair of users using a correlation measure, wherein the user-to-user similarity matrix contains pairs of values, wherein each pair is assigned to each entry in the user-item matrix;

assign a weighted score for each user and item according to the level of similarity to other users;

sort, in said database, all the items in the form of a catalog of items according to their weighted scores as calculated for the user and filtering out all items that have been already rated by said user; and

automatically present to the user, on the electronic device of the user, items from said catalog of items, rated according to their weighted scores, wherein presented items are fine-grained geo-tagged items that are recommended based on geo-information and user'"'"'s rating.

2. The recommender system according to claim 1, wherein the at least one electronic device is selected from the group consisting of:

personal computers;

mobile computers;

PDAs;

Cameras;

or

cellular phones.

3. The recommender system according to claim 1, wherein the recommendation server measures similarity by analyzing the geographic distance of the GeoTagged items and finds new geographical places and items related to those places that are of interest to the user.

4. The recommender system according to claim 1, wherein the recommendation server recommends items associated to close-by locations and items in similar locations elsewhere.

5. The recommender system according to claim 1, wherein the recommendation server includes a mechanism adapted to service requests from a user for computational and data storage resources.

6. The recommender system according to claim 1, wherein the database contains data and processes for analyzing the GeoTag information.

7. The recommender system according to claim 1, wherein the database contains a table for storing the feedbacks obtained from users regarding the various items.

8. The recommender system according to claim 1, wherein each pair includes a first value that represents the mean propagated rating and a second value that refers to the total propagated weights.

9. The recommender system according to claim 1, wherein each item is associated with Longitude and Latitude information for calculating the geographical distance between any pair of items by using database built-in functions.

10. The recommender system according to claim 1, wherein the recommendation server calculates a user-to-user similarity matrix for each pair of users according to the revised Pearson correlation:

11. The recommender system according to claim 1, wherein the rating is of a binary type.

12. The recommender system according to claim 1, wherein the rating is implicit or explicit.

13. The recommender system according to claim 1, wherein the rating is determined by observing the user'"'"'s behavior.

14. The recommender system according to claim 1, wherein the rating is a “

visited”

or a “

check-in”

event.

15. The recommender system according to claim 1, in which the time and date of each rating is recorded for recommending locations.

16. The recommender system according to claim 1, in which the time attribute is used to further distinct the ratings of the user, while calculating a recommendation.