Granted Patent B2

US 10,423,983 · App. 14/488,074 · Granted Sep 24, 2019

Abstract

Claims (80)

1. A method for use by at least one data processing device, the method comprising:

receiving, from a client device, targeting criteria that comprises at least a targeted behavior and an identification of a geographic region;

segmenting the geographic region identified by the targeting criteria into a three-dimensional grid that comprises a plurality of cells based on latitude data, longitude data, and altitude data, the plurality of cells including a cell that encompasses a location within the region, the cell comprising a cell identifier;

receiving behavioral information associated with at least a user, wherein the behavioral information includes location data that comprises at least device altitude data, and temporal data, the location data identifying the location encompassed by the cell;

calculating a behavior match metric for the cell based on the behavioral information projected onto the geographic region, the behavior match metric indicating a number of visits by the user to the location represented by the cell;

assigning the behavioral match metric to a key that comprises a tuple that comprises the cell identifier of the cell and a user identifier of the user;

retrieving feature data for the cell, the feature data indicating attributes of the location encompassed by the cell;

generating a labeled feature vector table based on the key and the attributes of the location encompassed by the cell;

training a model for predicting a conversion rate of the cell based on the labeled feature vector table, wherein the conversion rate provides a probability of the user performing the targeted behavior within the cell;

applying the model to the feature data to predict the conversion rate of the cell; and

presenting targeting information based on the conversion rate of the cell to the client device.

2. The method of claim 1 , wherein the cell is a first cell, and the identifying the targeting information based on the conversion rate of the first cell further comprises:

ranking the first cell among the plurality of cells based on the conversion rate; and

selecting the first cell based on the ranking of the first cell among the plurality of cells.

3. The method of claim 2 , further comprising providing latitude and longitude coordinates corresponding to each cell among the plurality of cells as the targeting information to an entity from which the targeting criteria were received.

4. The method of claim 2 , further comprising:

determining one or more unique identifiers of users observed in the first cell, in response to the selecting the first cell; and

providing the one or more unique identifiers as the targeting information to an entity from which the targeting criteria were received.

5. The method of claim 1 , wherein the identifying the targeting information based on the conversion rate of the cell further comprises:

calculating an aggregate conversion rate for the cell based on one or more conversion rates of the plurality of cells within a predefined distance from the cell;

selecting cells from among the plurality of cells with higher aggregate conversion rates than others; and

providing latitude and longitude coordinates corresponding to the selected cells as the targeting information to an entity from which the targeting criteria were received.

6. The method of claim 1 , wherein the behavioral information further includes survey data and demographic profile data associated with the multiple users.

7. The method of claim 6 , further comprising:

identifying a global user set from the multiple users based on at least one of the survey data or the demographic profile data;

wherein the global user set includes the multiple users when no criteria for identifying users based on the survey data or the demographic profile data is provided.

8. The method of claim 7 , further comprising:

identifying a behavior match user set from the global user set based on the survey data;

wherein the behavior match user set is the same as the global user set when no criteria for identifying the behavior match user set based on the survey data is provided.

9. The method of claim 8 , further comprising:

identifying a global visit set as a subset of all visits performed by users in the global user set; and

identifying a subset of the global visit set that matches the targeted behavior and was performed by a user in the behavior match user set as behavior match visit set.

10. The method of claim 9 , wherein calculating the behavior match metric for each cell based on the behavioral information farther comprises:

grouping the behavior match visit set by a user and by day;

for each user in the behavior match user set, calculating a total number of days on which the user performed the targeted behavior, wherein the total number of days corresponds to the behavior match metric.

11. The method of claim 1 , wherein labeling the feature data for the cell includes using the cell identifier of the cell to join a key-value pair having a tuple of the cell identifier and a user identifier as a key and a behavior match metric of the cell as a value with another key-value pair having the cell identifier as a key and the feature data as a value.

12. The method of claim 1 , further comprising:

generating the feature data for the cell, wherein the feature data includes:

nearby place features related to all places that are within a predefined distance from each cell;

place distance features related to minimum distances from the cell to each business identifier and each category identifier respectively that are within a predefined distance from the cell; and

demographic features for the cell obtained from census data.

13. The method of claim 1 , wherein training the model for predicting the conversion rate of the cell based on the set of the labeled feature data further comprises:

training a classifier using the set of the labeled feature data to predict a visit probability that provides an indication of a likelihood that a user observed in the cell would perform the targeted behavior; and

training a statistical model using the visit probability and an observed conversion rate aggregated across all users in the cell to predict the conversion rate for the cell.

14. The method of claim 13 , wherein applying the model to the feature data to predict the conversion rate of the cell further comprises:

applying the trained classifier on the feature data to predict a visit probability; and

applying the statistical model on the visit probability to predict the conversion rate for the cell.

15. The method of claim 14 , further comprising:

randomly sampling users to obtain the set of the labeled feature data from the feature data for training the model.

16. The method of claim 1 , wherein the targeted behavior includes visiting a place or performing an activity.

17. A system comprising:

memory;

at least one processor in communication with the memory and configured to execute a plurality of instructions stored in the memory to:

receive, from a client device, targeting criteria that comprises at least a targeted behavior and an identification of a geographic region;

segment the geographic region identified by the targeting criteria into a three-dimensional grid that comprises a plurality of cells based on latitude data, longitude data, and altitude data, the plurality of cells including a cell that encompasses a location within the region, the cell comprising a cell identifier;

receive behavioral information associated with at least a user, wherein the behavioral information includes location data that comprises at least device altitude data, and temporal data, the location data identifying the location encompassed by the cell;

calculate a behavior match metric for the cell based on the behavioral information projected onto the geographic region, the behavior match metric indicating a number of visits by the user to the location represented by the cell;

assign the behavioral match metric to a key that comprises a tuple that comprises the cell identifier of the cell and a user identifier of the user;

retrieve feature data for the cell, the feature data indicating attributes of the location encompassed by the cell;

generate a labeled feature vector table based on the key and the attributes of the location encompassed by the cell;

train a model for predicting a conversion rate of the cell based on the labeled feature vector table, the conversion rate providing a probability of the user performing the targeted behavior within the cell;

apply the model to the feature data to predict the conversion rate of the cell; and

present targeting information based on the conversion rate of the cell to the client device.

18. The system of claim 17 , wherein the targeting information includes at least one of:

latitude and longitude coordinates of the plurality of cells; or

one or more unique identifiers associated with users observed in the plurality of cells.

19. A non-transitory computer-readable medium storing computer-executable instructions that cause a machine to perform operations comprising:

receiving, from a client device, targeting criteria that comprises at least a targeted behavior and an identification of a geographic region;

segmenting the geographic region identified by the targeting criteria into a three-dimensional grid that comprises a plurality of cells based on latitude data, longitude data, and altitude data, the plurality of cells including a cell that encompasses a location within the region, the cell comprising a cell identifier;

receiving behavioral information associated with at least a user, wherein the behavioral information includes location data that comprises at least device altitude data, and temporal data, the location data identifying the location encompassed by the cell;

calculating a behavior match metric for the cell based on the behavioral information projected onto the geographic region, the behavior match metric indicating a number of visits by the user to the location represented by the cell;

assigning the behavioral match metric to a key that comprises a tuple that comprises the cell identifier of the cell and a user identifier of the user;

retrieving feature data for the cell, the feature data indicating attributes of the location encompassed by the cell;

generating a labeled feature vector table based on the key and the attributes of the location encompassed by the cell;

training a model for predicting a conversion rate of the cell based on the labeled feature vector table, wherein the conversion rate provides a probability of the user performing the targeted behavior within the cell;

applying the model to the feature data to predict the conversion rate of the cell; and

presenting targeting information based on the conversion rate of the cell to the client device.

20. The medium of claim 19 , wherein the cell is a first cell, and the identifying the targeting information based on the conversion rate of the first cell further comprises:

ranking the first cell among the plurality of cells based on the conversion rate; and

selecting the first cell based on the ranking of the first cell among the plurality of cells.

Assignments (9)

Cited By (3)

US 12,216,519 US 12,513,254 US 12,645,280