Granted Patent B2

US 8,019,777 · App. 12/795,419 · Granted Sep 13, 2011

Assignee: Nexify, Inc.

Abstract

Claims (64)

1. A method of producing and storing a content interest profile, said method comprising:

using a communications interface receiving from the user a request for service of one or more content items;

receiving and storing a unique identifier, URL content, time period, and IP address of the user;

using a processor device configured to operate as:

a user profile generator tracking digital content consumed by a user and user interaction with said digital content by performing steps of:

determining whether the user has been provided the unique identifier (ID);

if the user has not been provided the unique identifier, assigning the unique identifier to the user;

placing software code in the user's web site to read and record a URL content (uniform resource locator) of web pages or other address from which digital content can be retrieved;

storing the unique identifier in a database; and

using page-based JavaScript code to report the interaction in real-time via an application programming interface; and

constructing a detailed profile of the user's interests based on the digital content consumed and the interaction with the digital content;

using a processor device configured to operate as:

a content analysis and data mining engine performing collaborative filtering, comprising processing the digital content through data-mining, semantic filters, and content enrichment processing to extract topics, concepts, and entities related to each content item;

assigning a relevance score to the processed digital content, wherein said relevance score is based on frequency, content relevance, collaborative filtering scores, social tags, user actions, prior domain knowledge, a measure of recency, editorial rules, and a commercial value of the content;

storing the relevance score;

ranking recommendable content based on the relevance score;

using a processor device configured to operate as:

a content recommendation engine determining what digital content is likely to be of interest to the user by matching the user's interests from the profile built by the user profile generator with an available pool of content to find a closest match to the user's interest, factoring in relevance, recency of the digital content based on either its initial publication or subsequent updates, popularity, and an interest of other similar users.

2. The method of claim 1 wherein the assigning step comprises inserting JavaScript into the user's computer's memory to assign a cookie to the user.

3. The method of claim 1 further comprising a step of receiving and storing the user's longitude and latitude.

4. The method of claim 1 further comprising a step of receiving and storing the user's device type.

5. The method of claim 1 further comprising a step of receiving and storing the user's referrer.

6. The method of claim 1 further comprising analyzing subject matter of each digital content previously viewed by the user, using a plurality of content analysis processes to determine the meaning and substance of each content article.

7. The method of claim 6 wherein the content analysis processes comprise semantic analysis of content, software for entity extraction, knowledge maps and other methods of determining topical matter.

8. The method of claim 1 wherein the digital content viewed comprises text, audio, and video of graphical file.

9. The method of claim 1 wherein the digital content comprises a news story.

10. The method of claim 8 wherein where the digital content viewed comprises video or audio and the content is processed with speech recognition in order to generate ancillary text to be data mined and semantically analyzed.

11. The method of claim 9 further comprising dynamically and continuously updating each user's changes in interest based on the digital content viewed.

12. The method of claim 1 further comprising providing the user the ability to declare interest in a topic.

13. A method of applying statistical, probabilistic, and predictive methods to data contained in a user's profile in order to determine the likely interest of a user in content not yet viewed, said method comprising steps of:

using an information processing device for:

using a plurality of calculations to determine a pattern in a subject matter of articles previously viewed by determining similarities in substance between articles previously viewed and a set of articles not yet viewed, said determining comprising:

comparing a plurality of data points, wherein said data points comprise a content's author, source, time of creation, time of publication, and length;

wherein the length is measured in terms of character, word or paragraph count for text assets;

selecting content most likely to be of interest to the user by using a matching function that is performed based on recency of the content based on either its initial publication or subsequent updates; and

ranking recommendable assets based on a series of signals including the content relevance, collaborative filtering scores, social tags, user actions, prior domain knowledge, a measure of recency, editorial rules and the commercial value of the asset.

14. The method of claim 13 further comprising ranking assets not yet viewed according to a likelihood of interest for each user.

15. The method of claim 13 further comprising including an ability of a publisher to access a dashboard where statistical matches between user profiles and available content may be viewed at an aggregate level to provide a more comprehensive perspective of the audience content preferences.

16. A method for making personalized content recommendations to users, said method comprising:

using an information processing device for:

scoring each available content asset against a content preference of each user contained in a user profile;

applying statistical, probabilistic and predictive methods to data contained in a user's profile in order to determine a likely interest of a user in content not yet viewed; and

ranking recommendable content assets based on a series of signals including the content relevance, collaborative filtering scores, social tags, user actions, prior domain knowledge, a measure of recency, editorial rules and the commercial value of the asset.

17. The method of claim 16 further comprising a step of using a matching function to determine a highest match between the available content assets and a user's profile.

18. The method of claim 16 further comprising:

preparing a set of recommended content assets;

wherein the recommended content assets comprise an entire new page of content created solely for the user.

19. The method of claim 16 wherein the recommended content asset further comprises:

elements within a page.

20. The method of claim 16 wherein the recommended content asset further comprises:

an advertisement or a commercial good or service available for sale.

21. The method of claim 16 wherein the recommended content asset further comprises:

recommendations populating screens of a mobile application with data relevant to the user.

22. The method of claim 16 further comprising:

providing a system supporting the method in exchange for a fee to be paid by a content publisher, a vendor of goods and a provider of services,

wherein the fee comprises a fixed, recurring payment or a variable payment proportionate to certain other variables measuring the effectiveness of the recommendations.

23. The method of claim 1 wherein determining what digital content is likely to be of interest to the user comprises selecting at least one factor from a group consisting of:

evaluating the user's interest over different time periods and across different domains and media types;

balancing long and short term user interests;

time period of content;

value of certain topics or entities;

editorial recommendations; and

economic value of presenting a certain digital content over another.

24. The method of claim 1 wherein the content recommendation engine factors a success of different recommendation methods, such as collaborative filtering and contextual relevance, to learn from and adjust to the recommendation method that is most successful with a specific user.

Assignments (13)

Continuity (2)