Generating an audience similarity model

Abstract

The systems and methods described herein can determine the audience for content delivery. The server receives configuration input from the content-generating user to identify the target audience. The configuration input indicates the target audience, baseline audience, and special audience. The target audience includes the end users targeted by the content. The baseline audience includes a broad group of end users with specific characteristics (e.g., geography). The special audience includes a selected group of end users with specific characteristics to topics previously accessed by the end users, and the server correlates the baseline audience and special audience to identify the target audience of target users who access web pages on those topics and, for example, reside in a certain region. The server generates a ranked list of contextual terms applied by the content user and uses topical terms from all target users to predict whether a web page has a target audience.

Claims

1. A computer-based method for determining contextually relevant audiences for content distribution, A computer receives one or more configuration inputs via a content user's user interface, wherein the one or more configuration inputs indicate a target audience and one or more contextual terms. The steps include: using the computer to identify a set of target users associated with one or more contextual terms that define the target audience by cross-referencing a first group of end users of a special audience with a second group of end users of a background audience; The steps include: using the computer to identify a ranked list of contextual terms associated with each target end user of the target audience; A method comprising the steps of: training a classifier that predicts the probability of a web page having a similar audience by applying the classifier to a ranked list of contextual terms associated with the target audience.

2. The method according to claim 1, further comprising the step of applying the classifier to a plurality of topic terms of the web page in order to predict the likelihood of the web page being the similar audience of the web page, using the computer.

3. The method according to claim 1, further comprising the step of the computer generating the special audience based on user data of each end user among the first plurality of end users of the special audience, according to one or more configuration inputs.

4. The method according to claim 3, further comprising the step of updating the special audience in accordance with additional user data received from one or more client devices by the computer.

5. The method according to claim 1, further comprising the step of the computer selecting the background audience from a database according to the background features indicated by the one or more configuration inputs.

6. The method according to claim 5, wherein the step of selecting the background audience includes the step of the computer extracting a sample subset of user data records from the database of the second plurality of end users of the background audience.

7. The method according to claim 1, further comprising the step of determining multiple co-occurrence probabilities of multiple topic terms in multiple corpus web pages using the computer.

8. For a specific end user, the computer identifies one or more historical web pages accessed by that specific end user; The steps include: using the computer to identify multiple topic terms of one or more historical web pages accessed by the specific end user; The method according to claim 1, further comprising the step of updating the data record of the particular end user to include the plurality of topic terms using the computer.

9. The method according to claim 1, further comprising the step of transmitting a command from the computer to a client device to display the target audience via the user interface of the client device.

10. The steps include: receiving an availability list of multiple available web pages requesting bids from the bidding server via the computer; The method according to claim 1, further comprising the step of generating the probability of the similar audience for the available web page by applying the classifier to a plurality of topic terms of the available web page by the computer for each available web page in the bidding stream.

11. A system for determining the audience for contextually relevant content distribution, Receiving one or more configuration inputs through the user interface of a content user, wherein the one or more configuration inputs indicate a target audience and one or more contextual terms, By cross-referencing a first group of end users of a special audience with a second group of end users of a background audience, the set of target users associated with the one or more contextual terms that define the target audience is identified. Identifying a ranked list of contextual terms associated with each target end user of the aforementioned target audience, A system comprising a computer having at least one processor configured to train a classifier that predicts the probability of a web page being a similar audience by applying the classifier to a ranked list of contextual terms associated with the target audience.

12. The system according to claim 11, wherein the computer is further configured to apply the classifier to a plurality of topic terms of the web page in order to predict the likelihood of the web page being the similar audience of the web page.

13. The system according to claim 11, wherein the computer is further configured to generate the special audience based on user data of each end user among the first plurality of end users of the special audience, according to one or more configuration inputs.

14. The system according to claim 13, wherein the computer is further configured to update the special audience in accordance with additional user data received from one or more client devices.

15. The system according to claim 11, wherein the computer is further configured to select the background audience from a database according to the background characteristics indicated by the one or more configuration inputs.

16. The system according to claim 15, wherein, when selecting the background audience, the computer is further configured to extract a sample subset of user data records from the database of the second plurality of end users of the background audience.

17. The system according to claim 11, wherein the computer is further configured to determine multiple co-occurrence probabilities of multiple topic terms in multiple corpus web pages.

18. The aforementioned computer, For a specific end user, to identify one or more historical web pages accessed by that specific end user, Identifying multiple topic terms from one or more historical web pages accessed by the aforementioned specific end user, The system according to claim 11, further configured to update the data record of a particular end user in order to include the aforementioned multiple topic terms.

19. The system according to claim 11, wherein the computer is further configured to transmit commands to a client device for displaying the target audience via the user interface of the client device.

20. The aforementioned computer, The process involves receiving an availability list of multiple available web pages requesting bids from the bidding server, and The system according to claim 11, further configured to generate the probability of the similar audience for each available web page in the bidding stream by applying the classifier to a plurality of topic terms for the available web page.

Images