ONLINE BEHAVIORAL PREDICTOR

• US 20150278837A1
• Filed: 04/04/2014
• Published: 10/01/2015
• Est. Priority Date: 03/31/2014
• Status: Active Grant

• Current Assignee: LivePerson Incorporated
• Original Assignee: LivePerson Incorporated
• Inventors: LAHAV, Shlomo, RON, Ofer, SHIMON, David
• Granted Patent: US 11,386,442 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06Q 30/0204 Market segmentation

Abstract

In some embodiments, a set of user groups can be defined, with each group relating to a different webpage experience, user action, etc. Requests are assigned to one of the groups based on actual webpage presentation features and/or user actions. A group-specific model is generated for each group and translates user information to a preliminary result (e.g., a purchasing probability). A model combination includes a weighted combination of a set of available group-specific models. User information is processed using the model combination to generate a model result. The model result is evaluated to determine whether a requested webpage is to be customized in a particular manner and/or an opportunity is to be offered.

Claims

1. A computer-implemented method comprising:

accessing a set of group-specific models, wherein;

each group-specific model in the set of group-specific models is configured to generate a prediction of a same action variable;

the prediction of the same action variable is indicative of an interaction with a version of a webpage;

the version of the webpage is being provided in response to a request for the webpage;

the prediction of the same action variable is generated using information associated with an identifier in the request for the webpage;

the information characterizes a user associated with the identifier or characterizes a previous webpage interaction associated with the identifier;

the previous webpage interaction corresponds to an interaction or lack of interaction with the webpage or a different webpage;

each group-specific model in the set of group-specific models corresponds to a webpage feature indicating whether or which electronic opportunity is offered in response to the request for the webpage; and

each group-specific model in the set of group-specific models has been derived, at least in part, based on data characterizing interactions or lack of interaction with one or more webpages having the webpage feature;

determining an accuracy for each of one or more group-specific models of the set of group-specific models;

developing a model combination that includes a weighted combination of a plurality of the set of group-specific models, wherein;

the model combination is configured to generate a prediction of the same action variable; and

the model combination includes a plurality of dynamic weights determined based on the accuracy determined for each of the one or more group-specific models;

receiving a new request for the webpage, wherein the new request includes a new identifier;

determining a model result using the model combination and new information associated with the new identifier, wherein the model result includes a prediction of the same action variable;

using the model result to provide a response to the new request for the webpage;

determining a webpage configuration characteristic based on the model result;

generating or identifying a version of the webpage such that the generated or identified version of the webpage has the webpage configuration characteristic; and

transmitting the generated or identified version of the webpage to a device associated with the new request.

3. The method of claim 1, wherein the webpage feature is indicative of whether a chat invitation was or is to be electronically extended in association with the webpage.

4. The method of claim 1, wherein the model result includes a prediction as to whether an online purchase will be made during a session associated with the second request.

5. The method of claim 1, wherein, for a group-specific model in the set of group-specific models, the information includes data from a cookie associated with the request for the webpage.

6. The method of claim 1, wherein determining the model result includes:

determining, for each group-specific model in the plurality of the developed group-specific models, a preliminary model result using at least some of the new information and the group-specific model;

weighting each of the preliminary model results using a weight in the plurality of dynamic weights; and

combining the weighted preliminary model results.

8. The method of claim 1, wherein the model result includes a numeric value, and wherein determining the webpage configuration characteristic based on the model result includes:

comparing the model result to a threshold.

21. The method of claim 1, wherein the configuration characteristic indicates whether or which electronic opportunity is to be offered via the version of the webpage.

22. The method of claim 1, wherein the configuration characteristic indicates whether an electronic chat is to be offered via the version of the webpage.

23. The method of claim 1, further comprising:

identifying, subsequent to the transmission, an observation of the same action variable;

determining, for a group-specific model in the plurality of the developed group-specific models, a predicted model result using at least some of the new information and the group-specific model;

comparing the observation of the same action variable to the predicted model result; and

modifying a dynamic weight in the plurality of dynamic weights based on the comparison.

2. (canceled)

7. (canceled)

9. (canceled)

10. A system comprising:

one or more data processors; and

a non-transitory computer readable storage medium containing instructions which when executed on the one or more data processors, cause the one or more data processors to perform operations including;

accessing a set of group-specific models, wherein;

each group-specific model in the set of group-specific models is configured to generate a prediction of a same action variable;

the prediction of the same action variable is indicative of an interaction with a version of a webpage;

the version of the webpage is being provided in response to a request for the webpage;

the prediction of the same action variable is generated using information associated with an identifier in the request for the webpage;

the information characterizes a user associated with the identifier or characterizes a previous webpage interaction associated with the identifier;

the previous webpage interaction corresponds to an interaction or lack of interaction with the webpage or a different webpage;

each group-specific model in the set of group-specific models corresponds to a webpage feature indicating whether or which electronic opportunity is offered in response to the request for the webpage; and

each group-specific model in the set of group-specific models has been derived, at least in part, based on data characterizing interactions or lack of interaction with one or more webpages having the webpage feature;

determining an accuracy for each of one or more group-specific models of the set of group-specific models;

developing a model combination that includes a weighted combination of a plurality of the set of group-specific models, wherein;

the model combination is configured to generate a prediction of the same action variable; and

the model combination includes a plurality of dynamic weights determined based on the accuracy determined for each of the one or more group-specific models;

receiving a new request for the webpage, wherein the new request includes a new identifier;

determining a model result using the model combination and new information associated with the new identifier, wherein the model result includes a prediction of the same action variable;

using the model result to provide a response to the new request for the webpage;

determining a webpage configuration characteristic based on the model result;

generating or identifying a version of the webpage such that the generated or identified version of the webpage has the webpage configuration characteristic; and

transmitting the generated or identified version of the webpage to a device associated with the new request.

12. The system of claim 10, wherein the webpage feature is indicative of whether a chat invitation was or is to be electronically extended in association with the webpage.

13. The system of claim 10, wherein the model result includes a prediction as to whether an online purchase will be made during a session associated with the second request.

14. The system of claim 10, wherein, for a group-specific model in the set of group-specific models, the information includes data from a cookie associated with the request for the webpage.

15. The system of claim 10, wherein determining the model result includes:

determining, for each group-specific model in the plurality of the developed group-specific models, a preliminary model result using at least some of the new information and the group-specific model;

weighting each of the preliminary model results using a weight in the plurality of dynamic weights; and

combining the weighted preliminary model results.

17. The system of claim 10, wherein the model result includes a numeric value, and wherein determining the webpage configuration characteristic based on the model result includes:

comparing the model result to a threshold.

24. The system of claim 10, wherein the configuration characteristic indicates whether or which electronic opportunity is to be offered via the version of the webpage.

25. The system of claim 10, wherein the configuration characteristic indicates whether an electronic chat is to be offered via the version of the webpage.

26. The system of claim 10, wherein the operations further include:

identifying, subsequent to the transmission, an observation of the same action variable;

determining, for a group-specific model in the plurality of the developed group-specific models, a predicted model result using at least some of the new information and the group-specific model;

comparing the observation of the same action variable to the predicted model result; and

modifying a dynamic weight in the plurality of dynamic weights based on the comparison.

11. (canceled)

16. (canceled)

18. (canceled)

19. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform operations including:

accessing a set of group-specific models, wherein;

each group-specific model in the set of group-specific models is configured to generate a prediction of a same action variable;

the prediction of the same action variable is indicative of an interaction with a version of a webpage;

the version of the webpage is being provided in response to a request for the webpage;

the prediction of the same action variable is generated using information associated with an identifier in the request for the webpage;

the information characterizes a user associated with the identifier or characterizes a previous webpage interaction associated with the identifier;

the previous webpage interaction corresponds to an interaction or lack of interaction with the webpage or a different webpage;

each group-specific model in the set of group-specific models corresponds to a webpage feature indicating whether or which electronic opportunity is offered in response to the request for the webpage; and

each group-specific model in the set of group-specific models has been derived, at least in part, based on data characterizing interactions or lack of interaction with one or more webpages having the webpage feature;

determining an accuracy for each of one or more group-specific models of the set of group-specific models;

developing a model combination that includes a weighted combination of a plurality of the set of group-specific models, wherein;

the model combination is configured to generate a prediction of the same action variable; and

the model combination includes a plurality of dynamic weights determined based on the accuracy determined for each of the one or more group-specific models;

receiving a new request for the webpage, wherein the new request includes a new identifier;

determining a model result using the model combination and new information associated with the new identifier, wherein the model result includes a prediction of the same action variable;

using the model result to provide a response to the new request for the webpage;

determining a webpage configuration characteristic based on the model result;

generating or identifying a version of the webpage such that the generated or identified version of the webpage has the webpage configuration characteristic; and

facilitating a transmission of the generated or identified version of the webpage to a device associated with the new request.

27. The computer-program product of claim 19, wherein the model result includes a prediction as to whether an online purchase will be made during a session associated with the second request.

20. (canceled)