Online behavioral predictor
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.
1. A computer-implemented method comprising:
receiving, by a computing device, set of webpage requests for access to a webpage, the set of webpage requests being associated with a set of users;
generating, by the computing device, a group-specific model corresponding to a user group of the set of users, the user group including users that share one or more characteristics, wherein the group-specific model is generated by training the group-specific model using historical data associated with the user group, wherein the group-specific model correlates particular webpage customizations to predicted user actions associated with the webpage, and wherein the group-specific model is configured to generate a prediction of a likelihood that a user in the user group will complete an action when the webpage includes a particular webpage customization;
generating, by the computing device, a first version of the webpage for users of the user group, the first version of the webpage including one or more customizations to the webpage selected based on predictions of the group-specific model;
facilitating, by the computing device, a transmission of the first version of the webpage to at least one user of the user group;
receiving, by the computing device, indications of instances of user interaction with the first version of the webpage;
retraining, by the computing device, the group-specific model by updating one or more weights of the model based on a comparison between predictions generated by the group-specific model and the indications of instances of user interaction with the first version of the webpage;
receiving, by the computing device, a request for access to the webpage, wherein the request is associated with a particular user of the user group;
generating, in response to the request and by the computing device, a second version of the webpage, the second version of the webpage including one or more alternate customizations to the webpage selected based on predictions of the retrained group-specific model, wherein the one or more alternate customizations increase the likelihood that a user will execute the action on the webpage; and
facilitating a transmission, by the computing device, of the second version of the webpage to the particular user.
2. The computer-implemented method of claim 1 , wherein the set of webpage requests are received over a time interval defined as a training increment for the group-specific model.
3. The computer-implemented method of claim 1 , wherein a predetermined percentage of the set of webpage requests are control requests.
4. The computer-implemented method of claim 1 , wherein the set of webpage requests are selected based on user characteristics associated with the set of users.
5. The computer-implemented method of claim 1 , wherein the one or more customizations include a chat interface.
6. The computer-implemented method of claim 1 , wherein the group-specific model is retrained at predetermined time intervals.
7. The computer-implemented method of claim 1 , wherein the group-specific model is retrained after a predetermined quantity of instances of user interaction with the first version of the webpage is detected.
8. A system comprising:
one or more processors; and
a non-transitory computer-readable medium storing instructions that when executed by the one or more processors, cause the one or more processors to perform operations including:
receiving, by a computing device, set of webpage requests for access to a webpage, the set of webpage requests being associated with a set of users;
generating, by the computing device, a group-specific model corresponding to a user group of the set of users, the user group including users that share one or more characteristics, wherein the group-specific model is generated by training the group-specific model using historical data associated with the user group, wherein the group-specific model correlates particular webpage customizations to predicted user actions associated with the webpage, and wherein the group-specific model is configured to generate a prediction of a likelihood that a user in the user group will complete an action when the webpage includes a particular webpage customization;
generating, by the computing device, a first version of the webpage for users of the user group, the first version of the webpage including one or more customizations to the webpage selected based on predictions of the group-specific model;
facilitating, by the computing device, a transmission of the first version of the webpage to at least one user of the user group;
receiving, by the computing device, indications of instances of user interaction with the first version of the webpage;
retraining, by the computing device, the group-specific model by updating one or more weights of the model based on a comparison between predictions generated by the group-specific model and the indications of instances of user interaction with the first version of the webpage;
receiving, by the computing device, a request for access to the webpage, wherein the request is associated with a particular user of the user group;
generating, in response to the request and by the computing device, a second version of the webpage, the second version of the webpage including one or more alternate customizations to the webpage selected based on predictions of the retrained group-specific model, wherein the one or more alternate customizations increase the likelihood that a user will execute the action on the webpage; and
facilitating a transmission, by the computing device, of the second version of the webpage to the particular user.
9. The system of claim 8 , wherein the set of webpage requests are received over a time interval defined as a training increment for the group-specific model.
10. The system of claim 8 , wherein a predetermined percentage of the set of webpage requests are control requests.
11. The system of claim 8 , wherein the set of webpage requests are selected based on user characteristics associated with the set of users.
12. The system of claim 8 , wherein the one or more customizations include a chat interface.
13. The system of claim 8 , wherein the group-specific model is retrained at predetermined time intervals.
14. The system of claim 8 , wherein the group-specific model is retrained after a predetermined quantity of instances of user interaction with the first version of the webpage is detected.
15. A non-transitory computer-readable medium storing instructions that when executed by one or more processors, cause the one or more processors to perform operations including:
receiving, by a computing device, set of webpage requests for access to a webpage, the set of webpage requests being associated with a set of users;
generating, by the computing device, a group-specific model corresponding to a user group of the set of users, the user group including users that share one or more characteristics, wherein the group-specific model is generated by training the group-specific model using historical data associated with the user group, wherein the group-specific model correlates particular webpage customizations to predicted user actions associated with the webpage, and wherein the group-specific model is configured to generate a prediction of a likelihood that a user in the user group will complete an action when the webpage includes a particular webpage customization;
generating, by the computing device, a first version of the webpage for users of the user group, the first version of the webpage including one or more customizations to the webpage selected based on predictions of the group-specific model;
facilitating, by the computing device, a transmission of the first version of the webpage to at least one user of the user group;
receiving, by the computing device, indications of instances of user interaction with the first version of the webpage;
retraining, by the computing device, the group-specific model by updating one or more weights of the model based on a comparison between predictions generated by the group-specific model and the indications of instances of user interaction with the first version of the webpage;
receiving, by the computing device, a request for access to the webpage, wherein the request is associated with a particular user of the user group;
generating, in response to the request and by the computing device, a second version of the webpage, the second version of the webpage including one or more alternate customizations to the webpage selected based on predictions of the retrained group-specific model, wherein the one or more alternate customizations increase the likelihood that a user will execute the action on the webpage; and
facilitating a transmission, by the computing device, of the second version of the webpage to the particular user.
16. The non-transitory computer-readable medium of claim 15 , wherein the set of webpage requests are received over a time interval defined as a training increment for the group-specific model.
17. The non-transitory computer-readable medium of claim 15 , wherein a predetermined percentage of the set of webpage requests are control requests.
18. The non-transitory computer-readable medium of claim 15 , wherein the set of webpage requests are selected based on user characteristics associated with the set of users.
19. The non-transitory computer-readable medium of claim 15 , wherein the one or more customizations include a chat interface.
20. The non-transitory computer-readable medium of claim 15 , wherein the group-specific model is retrained at predetermined time intervals.