Selecting location-based content

• US 8,874,144 B1
• Filed: 12/20/2013
• Issued: 10/28/2014
• Est. Priority Date: 10/07/2013
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Liu, Yifang, Wu, Ping
• Primary Examiner(s): TRAN, CONGVAN
• Application Number:
  US14/137,480
• Time in Patent Office: 312 Days
• Field of Search: 455/456.1, 455/456.3, 455/456.6, 455/457, 455/566, 455/145, 715/273, 715/703, 715/761, 715/828, 715/864
• US Class Current: 455/456.3
• CPC Class Codes: H04W 4/02 Services making use of location information
• CPC Class Codes: H04W 4/029 Location-based management or tracking services

Abstract

Systems and methods include retrieving data indicative of a sequence of content that were previously selected for presentation by a mobile device based in part on physical locations of the mobile device. The sequence of selected content may be used with the current location of the mobile device to determine a likelihood score for a piece of content. Based on the likelihood score, the piece of content may be selected and provided to the mobile device for presentation.

Claims

1. A method for selecting location-based content comprising:

providing, for display on a user device, a first plurality of pieces of third-party content based on a corresponding plurality of locations of the user device, a first location of the corresponding plurality of locations different than a second location of the corresponding plurality of locations;

determining a sequence of third-party content indicating a temporal order in which the first plurality of pieces of third-party content are provided for display on the user device, the sequence comprising topical information and temporal information for each of the first plurality of pieces of third-party content;

identifying, in a data structure stored in a memory element, based on the topical and temporal information of the sequence, historical sequences of third-party content corresponding to the sequence, the historical sequences comprising topical and temporal information corresponding to each of a second plurality of pieces of third-party content in addition to the topical and temporal information corresponding to each the first plurality of pieces of third-party content;

generating likelihood scores for the second plurality of pieces of third-party content, the likelihood scores generated using a machine learning model that determines, based on the historical sequences, a likelihood of an interaction occurring with the second plurality of pieces of third-party content at the user device;

selecting, based on a comparison of likelihood scores, an additional piece of third-party content different from the first plurality of pieces of third-party content; and

providing, responsive to the sequence of the first plurality of pieces of third-party content provided for display, the additional piece of third-party content for display on the user device.

2. The method of claim 1, wherein topical information of the additional piece of third-party content is different from the topical information of each of the first plurality of pieces of third-party content.

3. The method of claim 1, wherein the machine learning model is a logistic regression model.

4. The method of claim 1, further comprising:

receiving a request for a piece of third-party content for display on the user device;

determining a current location of the user device; and

generating the likelihood scores using the current location of the user device.

5. The method of claim 1, further comprising:

receiving a request for a piece of third-party content for display on the user device;

determining a location trajectory of the user device based on the sequence;

identifying a proxy location for the user device based on the location trajectory; and

generating the likelihood scores using the proxy location of the user device.

6. The method of claim 1, further comprising:

determining the location trajectory of the user device based on a comparison of the sequence with the historical sequences.

7. The method of claim 1, further comprising:

selecting the additional piece of content corresponding to a highest likelihood score of the second plurality of pieces of content.

8. The method of claim 1, further comprising:

determining topical information corresponding to a highest likelihood score; and

selecting, based on the topical information, the additional piece of third-party content.

9. The method of claim 1, further comprising:

using the sequence to train the machine learning model.

10. The method of claim 9, further comprising:

retrieving data indicative of a second sequence of third-party content that were determined to be relevant to the user device based on the plurality of locations of the user device but were not selected for display on the user device; and

using the second sequence of third party-content to train the machine learning model.

11. A system for selecting location-based content comprising one or more processors configured to:

provide, for display on a user device, a first plurality of pieces of third-party content based on a corresponding plurality of locations of the user device, a first location of the

corresponding plurality of locations different than a second location of the corresponding plurality of locations;

determine a sequence of third-party content indicating a temporal order in which the first plurality of pieces of third-party content are provided for display on the user device, the sequence comprising topical information and temporal information for each of the first plurality of pieces of third-party content;

identify, in a data structure stored in a memory element, based on the topical and temporal information of the sequence, historical sequences of third-party content corresponding to the sequence, the historical sequences comprising topical and temporal information corresponding to each of a second plurality of pieces of third-party content in addition to the topical and temporal information corresponding to each the first plurality of pieces of third-party content;

generate likelihood scores for the second plurality of pieces of third-party content, the likelihood scores generated using a machine learning model that determines, based on the historical sequences, a likelihood of an interaction occurring with the second plurality of pieces of third-party content at the user device;

select, based on a comparison of likelihood scores, an additional piece of third-party content different from the first plurality of pieces of third-party; and

provide, responsive to the sequence of the first plurality of pieces of third-party content provided for display, the additional piece of third-party content for display on the user device.

12. The system of claim 11, wherein topical information of the additional piece of third-party content is different from the topical information of each of the first plurality of pieces of third-party content.

13. The system of claim 11, wherein the machine learning model is a logistic regression model.

14. The system of claim 11, wherein the one or more processors are further configured to:

receive a request for a piece of third-party content for display on the user device;

determine a current location of the user device; and

generate the likelihood scores using the current location of the user device.

15. The system of claim 11, wherein the one or more processors are further configured to:

receive a request for a piece of third-party content for display on the user device;

determine a location trajectory of the user device based on the sequence;

identify a proxy location for the user device based on the location trajectory; and

generate the likelihood scores using the proxy location of the user device.

16. The system of claim 11, wherein the one or more processors are further configured to:

determine the location trajectory of the user device based on a comparison of the sequence with the historical sequences.

17. The system of claim 11, wherein the one or more processors are further configured to:

select the additional piece of content corresponding to a highest likelihood score of the second plurality of pieces of content.

18. The system of claim 11, wherein the one or more processors are further configured to:

determine topical information corresponding to a highest likelihood score; and

select, based on the topical information, the additional piece of third-party content.

19. The system of claim 11, wherein the one or more processors are further configured to:

use the sequence to train the machine learning model.

20. The system of claim 19, wherein the one or more processors are further configured to:

retrieve data indicative of a second sequence of third-party content that were determined to be relevant to the user device based on the plurality of locations of the user device but were not selected for display on the user device; and

use the second sequence of third party-content to train the machine learning model.