Process for improving the quality of documents using semantic analysis

• US 7,693,705 B1
• Filed: 02/16/2005
• Issued: 04/06/2010
• Est. Priority Date: 02/16/2005
• Status: Expired due to Fees

• Current Assignee: Patrick William Jamieson
• Original Assignee: Patrick William Jamieson
• Inventors: Jamieson, Patrick William
• Primary Examiner(s): Hudspeth; David R
• Assistant Examiner(s): Spooner; Lamont M
• Application Number:
  US11/058,878
• Time in Patent Office: 1,875 Days
• Field of Search: 704/9, 704/270.1, 704/257, 704/4, 704/231, 704/270, 704/1, 704/10, 707/2, 707/3, 707/6
• US Class Current: 704/9
• CPC Class Codes: G06F 40/103 Formatting, i.e. changing of presentation of documents automatic justification G06F40/189; automatic line break hyphenation G06F40/191
• CPC Class Codes: G06F 40/20 Natural language analysis semantic analysis of natural language G06F40/30

Abstract

The present invention provides a new method for returning context specific questions during the process of authoring a document or report. The questions alert an author of possible omissions and vague or general statements. The method employs techniques in semantic analysis including corpus linguistics and sentential logic. The method takes as input sentences created by the document author, and looks up semantic propositions connected to these sentences, which are in the document'"'"'s knowledge domain, and questions linked to these propositions or other related propositions. The questions are sorted based on priority determined by a knowledge engineer. Additional methods are used to suppress redundant or nonspecific questions to increase their usefulness to the document author. The resulting question list aids the author in creating a high quality document.

Claims

1. A method for creating context specific document questions, comprising:

(a) obtaining an input sentence, and

(b) a computer processor retrieving said sentence from a database of known sentences in a knowledge domain created by segmenting at least a portion of a document corpus into individual sentences and,

(c) said processor retrieving corresponding sentential semantic proposition(s) which represent the entire meaning of said sentence from a semantic mapping table created by domain experts through a process of semantically annotating entire sentences from said corpus to sentential propositions, and

(d) said processor retrieving corresponding question(s) to said sentential semantic proposition(s).

2. The method according to claim 1 wherein a collection of unique questions are returned to a user for all the sentences in the document.

3. The method according to claim 1 wherein the input sentence is produced as the result of speech recognition or word processing.

4. The method according claim 1 wherein questions that have been given a higher priority by domain experts are displayed with a different visual indication from questions of lower priority such as a different color, font, size, highlighting, label, or any combination.

5. The method according to claim 1 wherein the return of a question is suppressed based on sentential proposition(s) in the document and question suppression proposition(s) linked to said question.

6. The method according to claim 1 wherein a link to reference information is made to one or more question(s) which supplies a user with either background information on the question, or specific suggestions to incorporate into said user document.

7. The method according to claim 1 where a user sends input sentences over a computer network to a remote question selector and questions are returned over a computer network to said user.

8. The method according to claim 1 where a question is returned or displayed with the attribute, “

ANSWERED”

, or a similar attribute if one or more sentential proposition(s) in the document directly answer the said question.

9. The method according to claim 1 where questions are returned to a user in near real-time as sentences are added to a document.

10. The method according to claim 1 where questions are returned to a user after the entire document is created.

11. A method for creating context specific document questions, comprising:

(a) obtaining an input sentence, and

(b) a computer processor retrieving said sentence from a database of known sentences in a knowledge domain created by segmenting at least a portion of a document corpus into individual sentences and,

(c) said processor retrieving corresponding sentential semantic proposition(s) which represent the entire meaning of said sentence created by domain experts through a process of semantically annotating entire sentences from said corpus to sentential propositions, and

(d) said processor retrieving related sentential semantic proposition(s) to said sentential semantic proposition(s) in a co-occurrence table generated by analyzing said corpus, and

(e) said processor retrieving corresponding question(s) to said related sentential semantic proposition(s).

12. The method according to claim 11 wherein a collection of unique questions are returned to a user for all the sentences in the document.

13. The method according to claim 11 wherein the input sentence is produced as the result of speech recognition or word processing.

14. The method according claim 11 wherein questions that have been given a higher priority by domain experts are displayed with a different visual indication from questions of lower priority such as a different color, font, size, highlighting, label, or any combination.

15. The method according to claim 11 wherein the return of a question is suppressed based on sentential proposition(s) in the document and question suppression proposition(s) linked to said question.

16. The method according to claim 11 wherein a link to reference information is made to one or more question(s) which supplies the user with either background information on the question, or specific suggestions to incorporate into said user document.