METHODS AND APPARATUS FOR CLASSIFYING CONTENT

• US 20130212109A1
• Filed: 08/16/2012
• Published: 08/15/2013
• Est. Priority Date: 08/18/2011
• Status: Active Grant

• Current Assignee: Rally Health, Inc (UnitedHealth Group Incorporated)
• Original Assignee: Rally Health, Inc.
• Inventors: EVANCICH, Nicholas H., BOHAN, Patrick, VERSTANDIG, Grant
• Granted Patent: US 8,793,254 B2
• Time in Patent Office: Days
• US Class Current: 707/740
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records

Abstract

In some embodiments, a non-transitory processor-readable medium includes code to cause a processor to send a signal representing a first question and a set of pictogram answers associated with the first question and a second question, different from the first question, and a set of pictogram answers associated with the second question. The first question and the second question can define a health-related survey such as a health-risk assessment. The non-transitory processor-readable medium includes code to receive a user selection of a pictogram answer associated with the first question and receive a user selection of a pictogram answer associated with the second question. The non-transitory processor-readable medium includes code to define a health-related user profile based on the user selection to the first question and the second question.

Claims

1. A non-transitory processor-readable medium storing code representing instructions to be executed by a processor, the code comprising code to cause the processor to:

receive at least a portion of a first natural language document defining a term and having a set of keywords;

receive at least a portion of a second natural language document defining the term and having a set of keywords different from the set of keywords from the first natural language document;

automatically define a training set based on the set of keywords from the first natural language document and the set of keywords from the second natural language document;

categorize a third natural language document as related to the term based on the training set; and

send an indication that the third natural language document is related to the term.

2. The non-transitory processor-readable medium of claim 1, wherein:

the code to define includes code to identify a subset of keywords included in the set of keywords from the first natural language document and included in the set of keywords from the second natural language document, the training set including the subset of keywords; and

the code to categorize includes code to categorize the third natural language document as related to the term when usage of keywords from the subset of keywords within the third natural language document exceeds a threshold.

3. The non-transitory processor-readable medium of claim 1, further comprising code to cause the processor to:

receive at least a portion of each message from a plurality of messages posted by a plurality of users on a website and associated with the term;

identify a keyword absent from the training set based on the plurality of messages when usage of the keyword within the plurality of messages exceeds a threshold; and

modify the training set to include the keyword in response to the identifying.

4. The non-transitory processor-readable medium of claim 1, further comprising code to cause the processor to:

receive at least a portion of each message from a plurality of messages posted by a plurality of users on a website and associated with the term;

identify a keyword absent from the training set based on the plurality of messages;

modify the training set to include the keyword in response to the identifying to define a modified training set;

categorize a fourth natural language document as related to the term based on the modified training set; and

send an indication that the fourth natural language document is related to the term.

5. The non-transitory processor-readable medium of claim 1, wherein:

the term is a medical condition;

the first natural language document is associated with a medical website, the term is defined within the first natural language document using the set of keywords from the first natural language document; and

the second natural language document is associated with a medical website different from the medical website associated with the first natural language document, the term is defined within the second natural language document using the set of keywords from the second natural language document.

6. The non-transitory processor-readable medium of claim 1, wherein:

the portion of the first natural language document is received from a crawl engine that searched a website;

the portion of the second natural language document is received from the crawl engine that searched a website different from the website for the first natural language document.

7. The non-transitory processor-readable medium of claim 1, wherein the term is a first term, the training set is for the first term, the non-transitory processor-readable medium further comprising code to cause the processor to:

define a training set for a second term based on a set of keywords from a third natural language document defining the second term and a set of keywords from a fourth natural language document defining the second term;

define a training set for a third term based on the training set for the first term and the training set for the second term, the third term including the first term and the second term;

categorize a fifth natural language document as related to the third term based on the training set for the third term; and

send an indication that the fifth natural language document is related to the third term.

8. The non-transitory processor-readable medium of claim 1, further comprising code to cause the processor to:

receive an indication of an interest of a user, the interest corresponding to the term; and

send the indication of the third natural language document in response to the indication of the interest of the user.

9. A non-transitory processor-readable medium storing code representing instructions to be executed by a processor, the code comprising code to cause the processor to:

categorize a first natural language document as related to a term based on a training set;

send an indication that the first natural language document is related to the term;

receive at least a portion of each message from a plurality of messages posted by a plurality of users on a website and associated with the term;

identify a keyword absent from the training set based on the plurality of messages when usage of the keyword within the plurality of messages exceeds a threshold;

modify the training set to include the keyword in response to the identifying to define a modified training set;

categorize a second natural language document as related to the term based on the modified training set; and

send an indication that the second natural language document is related to the term.

10. The non-transitory processor-readable medium of claim 9, further comprising code to cause the processor to:

receive at least a portion of a third natural language document defining the term and having a set of keywords;

receive at least a portion of a fourth natural language document defining the term and having a set of keywords different from the set of keywords from the third natural language document;

automatically define the training set based on the set of keywords from the third natural language document and the set of keywords from the fourth natural language document.

11. The non-transitory processor-readable medium of claim 9, further comprising code to cause the processor to:

receive at least a portion of a third natural language document defining the term and having a set of keywords, the portion of the third natural language document is received from a crawl engine that searched a website;

receive at least a portion of a fourth natural language document defining the term having a set of keywords different from the set of keywords from the third natural language document, the portion of the fourth natural language document is received from the crawl engine that searched a website different from the website for the third natural language document;

automatically define the training set based on the set of keywords from the third natural language document and the set of keywords from the fourth natural language document.

12. The non-transitory processor-readable medium of claim 9, wherein:

the term is a medical condition;

the first natural language document is associated with a website different from the website associated with the plurality of messages; and

the second natural language document is associated with a website different from the website associated with the first natural language document and different from the website associated with the plurality of messages.

13. The non-transitory processor-readable medium of claim 9, wherein the term is a first term, the non-transitory processor-readable medium further comprising code to cause the processor to:

define a training set for a second term based on a set of keywords from a third natural language document defining the second term and a set of keywords from a fourth natural language document defining the second term;

define a training set for a third term based on the modified training set for the first term and the training set for the second term, the third term is associated with the first term and the second term;

categorize a fifth natural language document as related to the third term based on the training set for the third term; and

send an indication that the fifth natural language document is related to the third term.

14. The non-transitory processor-readable medium of claim 9, further comprising code to cause the processor to:

receive an indication of an interest of a user, the interest corresponding to the term; and

send the indication of the second natural language document in response to the indication of the interest of the user.

15. A non-transitory processor-readable medium storing code representing instructions to be executed by a processor, the code comprising code to cause the processor to:

define a training set for a first term based on a set of keywords from a first natural language document defining the first term and a set of keywords from a second natural language document defining the first term;

define a training set for a second term based on a set of keywords from a third natural language document defining the second term and a set of keywords from a fourth natural language document defining the second term;

define a training set for a third term based on the training set for the first term and the training set for the second term, the third term is associated with the first term and the second term;

categorize a fifth natural language document as related to the third term based on the training set for the third term; and

send an indication that the fifth natural language document is related to the third term.

16. The non-transitory processor-readable medium of claim 15, wherein:

the first term is a medical condition;

the first natural language document is associated with a medical website, the first term is defined within the first natural language document using the set of keywords from the first natural language document; and

the second natural language document is associated with a medical website different from the medical website associated with the first natural language document, the first term is defined within the second natural language document using the set of keywords from the second natural language document.

17. The non-transitory processor-readable medium of claim 15, wherein:

the first term is a first medical condition;

the second term is a second medical condition; and

the third term is a third medical condition, the third term includes the first term and the second term.

18. The non-transitory processor-readable medium of claim 15, wherein:

the first term is a first medical condition;

the second term is a second medical condition; and

the third term is a third medical condition, the training set for the third term includes at least a portion of the training set for the first term and at least a portion of the training set for the second term.

19. The non-transitory process-readable medium of claim 15, wherein:

the set of keywords from the first natural language document, the set of keywords from the second natural language document, the set of keywords from the third natural language document and the set of keywords from the fourth natural language document each is produced by crawling a website having the first natural language document, a website having the second natural language document, a website having the third natural language document and a website having the fourth natural language document.

20. The non-transitory process-readable medium of claim 15, wherein:

the code to define the training set for the third term includes selecting a subset of keywords for the third term from a set of keywords included in the training set for the first term and from a set of keywords included in the training set for the second term.

21. The non-transitory process-readable medium of claim 15, wherein:

the code to define the training set for the first term includes code to identify a subset of keywords for the first term included in the set of keywords from the first natural language document and included in the set of keywords from the second natural language document, the training set for the first term including the subset of keywords for the first term;

the code to define the training set for the second term includes code to identify a subset of keywords for the second term included in the set of keywords from the third natural language document and included in the set of keywords from the fourth natural language document, the training set for the second term including the subset of keywords for the second term;

the code to define the training set for the third term includes selecting a subset of keywords for the third term from the subset of keywords for the first term included in for the training set for the first term and from the subset of keywords for the second term included in the training set for the second term; and

the code to categorize includes code to categorize the fifth natural language document as related to the third term when usage of keywords from the subset of keywords for the third term within the fifth natural language document exceeds a threshold.

22. The non-transitory processor-readable medium of claim 15, further comprising code to cause the processor to:

receive an indication of an interest of a user, the interest corresponding to the third term; and

send the indication of the fifth natural language document in response to the indication of the interest of the user.