Advertisement placement method and system using semantic analysis

• US 20050216516A1
• Filed: 05/11/2005
• Published: 09/29/2005
• Est. Priority Date: 05/02/2000
• Status: Active Grant

• Current Assignee: TextWise Company, LLC
• Original Assignee: TextWise Company, LLC
• Inventors: Calistri-Yeh, Randall Jay, Howatt, Mary Louise, Mah, Clinton P., Ruan, Wen, Forrester, Maurice J., McKenna, Mary E.
• Granted Patent: US 7,912,868 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3347 using vector based model
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

An advertisement placement method and system for relating an advertisement to a dataset based on a trainable semantic vector (TSV) associated with the dataset and respective semantic representations of the advertisements. The trainable semantic vector associated with the dataset is generated based on at least one data point included in the dataset and known relationships between predetermined data points and predetermined categories. A comparison process is performed to determine a similarity between the trainable semantic vector associated with the dataset and the semantic representation associated with each of the plurality of advertisements. The system selectively relates one or more of the advertisements with the dataset based on a result of the comparison process. The selected advertisement or advertisements may be displayed with the dataset.

Claims

1. A method for selectively relating a first dataset to a second dataset, the method comprising the steps of:

accessing a semantic representation associated with the first dataset and a semantic representation associated with the second dataset, wherein at least one of the semantic representation associated with the first dataset and the semantic representation associated with the second dataset is a trainable semantic vector generated based on at least one data point included in a respective dataset and known relationships between predetermined data points and predetermined categories;

determining a similarity between the semantic representation associated with the first dataset and the semantic representation associated with the second dataset; and

selectively relating the first dataset to the second dataset based on a result of the determining step.

2. The method of claim 1, wherein one of the first dataset and the second dataset is an advertisement, and the other one of the first dataset and the second dataset is one or more documents, web pages, e-mails, RSS news feeds, data streams, broadcast data or information related to a user;

or a portion or a combination of one or more documents, web pages, e-mails, RSS news feeds, data streams, broadcast data or information related to a user.

3. The method of claim 1, wherein one of the first dataset and the second dataset is a portion of a document, a web page, an e-mail, RSS news feeds, a data stream, broadcast data or information related to a user, and the other one of the first dataset and the second dataset is at least one background data source.

4. The method of claim 1 further comprising the step of selectively conveying the first dataset along with the second dataset based on a result of the selectively relating step.

5. The method of claim 4, wherein the first dataset is conveyed by display or a link accessible from the second dataset.

6. The method of claim 1, wherein the at least one data point includes at least one of a word, a phrase, a character string, a pre-assigned keyword, a sub-dataset, meta information and information retrieved based on a link included in the respective dataset.

7. The method of claim 1, wherein the trainable semantic vector is pre-calculated and included in the respective dataset.

8. The method of claim 1, wherein the trainable semantic vector is dynamically generated.

9. The method of claim 1, wherein the predetermined data points include at least one language.

10. The method of claim 1, wherein the trainable semantic vector associated with the respective dataset is generated further based on information related to at least one user or at least one dataset linked to said dataset.

11. The method of claim 10, wherein the information related to the at least one user includes at least one of documents previously viewed, previous search requests, user preferences and personal information.

12. The method of claim 1, wherein the relating step comprises the steps of:

accessing a predetermined threshold; and

relating the first dataset to the second dataset only when the similarity between the semantic representation associated with the first dataset and the semantic representation of the second dataset exceeds the predetermined threshold.

13. The method of claim 1, wherein the trainable semantic vector is reduced by selectively setting a value of at least one selected vector entry to zero.

14. The method of claim 13, wherein the at least one selected vector entry has a value less than a predetermined threshold.

15. The method of claim 13, wherein the at least one selected vector entry has a value less than the Nth largest value and where N is a predetermined constant.

16. The method of claim 1, wherein the predetermined categories include at least one geographic category.

17. The method of claim 1, wherein the semantic representation associated with one of the first dataset and the second dataset includes information from a pre-assigned vector derived based on preferred categories selected from the predetermined categories.

18. The method of claim 1 further comprising the step of increasing a weight of at least one entry in the semantic trainable vector corresponding to at least one of the predetermined categories.

19. An advertisement placement system for relating one of a plurality of advertisements to a dataset, the system comprising:

a data processor configured to process data; and

a data storage system configured to store instructions which, upon execution by the data processor, controls the data processor to perform the steps of;

accessing a semantic representation associated with a first dataset and a semantic representation associated with a second dataset, wherein at least one of the semantic representation associated with the first dataset and the semantic representation associated with the second dataset is a trainable semantic vector generated based on at least one data point included in a respective dataset and known relationships between predetermined data points and predetermined categories;

determining a similarity between the semantic representation associated with the first dataset and the semantic representation associated with the second dataset; and

selectively relating the first dataset to the second dataset based on a result of the determining step.

20. A machine-readable medium carrying instructions which, upon execution of a data processing system, controls the data processing system to perform the machine-implemented steps of:

accessing a semantic representation associated with a first dataset and a semantic representation associated with a second dataset, wherein at least one of the semantic representation associated with the first dataset and the semantic representation associated with the second dataset is a trainable semantic vector generated based on at least one data point included in a respective dataset and known relationships between predetermined data points and predetermined categories;

determining a similarity between the semantic representation associated with the first dataset and the semantic representation associated with the second dataset; and

selectively relating the first dataset to the second dataset based on a result of the determining step.