Calculating Valence Of Expressions Within Documents For Searching A Document Index

• US 20090077069A1
• Filed: 08/29/2008
• Published: 03/19/2009
• Est. Priority Date: 08/31/2007
• Status: Active Grant

• Current Assignee: Zhigu Holdings Limited
• Original Assignee: Powerset Inc (Microsoft Corporation)
• Inventors: Van den Berg, Martin Henk, Pell, Barney, Polanyi, Livia
• Granted Patent: US 8,346,756 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/30 of unstructured textual data document management systems G06F16/93
• CPC Class Codes: G06F 16/319 Inverted lists

Abstract

Tools and techniques related to calculating valence of expressions within documents. These tools may provide methods that include receiving input documents for processing, and extracting expressions from the documents for valence analysis, with scope relationships occurring between terms contained in the expressions. The methods may calculate calculating valences of the expressions, based on the scope relationships between terms in the expressions.

Claims

1. At least one computer-readable storage medium having computer-executable instructions stored thereon which, when executed by a computer system, cause the computer system to perform a method comprising:

receiving at least one document for processing;

extracting at least one expression from the document for valence analysis, wherein a scope relationship occurs between at least two terms contained in the expression; and

calculating a valence of the expression based on the scope relationship between terms as defined in the expression.

2. The storage medium of claim 1, wherein the instructions for calculating a valence include instructions for calculating a valence of a discourse within the document that includes the expression.

3. The storage medium of claim 1, wherein the instructions for calculating a valence include instructions for calculating a valence of a paragraph within the document that includes the expression.

4. The storage medium of claim 1, wherein the instructions for calculating a valence include instructions for identifying at least one term within the expression that is within a scope of at least a second term within the expression, for calculating respective valence values associated with at least the first and second terms, and for computing a valence of the expression by analyzing a semantic relationship between the first and second terms.

5. The storage medium of claim 1, wherein the instructions for extracting at least one expression include instructions for extracting an expression that represents at least one reported speech event, reported thought event, reported belief event, reported opinion event, or reported attitude event.

6. The storage medium of claim 1, wherein the instructions for extracting at least one expression include instructions for extracting an expression that represents at least one direct speech event.

7. The storage medium of claim 1, wherein the instructions for extracting at least one expression include instructions for extracting an expression that represents at least one indirect speech event.

8. The storage medium of claim 7, wherein the reported speech event represents an embedded speech event, in which a speaker conveys information, and represents a dominating speech event, in which an author reports on the embedded speech event, and expresses sentiment toward the information or toward the speaker.

9. The storage medium of claim 8, wherein the instructions for calculating a valence include instructions for calculating a valence when the speaker expresses a sentiment toward the information.

10. The storage medium of claim 1, further comprising instructions for indexing the expression for searching by valence.

11. The storage medium of claim 1, wherein the instructions for calculating a valence include instructions for calculating a valence of an expression to account for metaphoric use of at least one term included within the expression.

12. The storage medium of claim 1, wherein the expression represents a first instance of quoted speech, further comprising instructions for extracting at least a second expression from the document, wherein the second expression represents a second instance of quoted speech that includes the first instance of quoted speech, and further comprising instructions for repeating the method for at least the second expression.

13. At least one computer-readable storage medium having computer-executable instructions stored thereon which, when executed by a computer system, cause the computer system to calculate a valence of at least one expression contained within an input document using a method comprising:

calculating a representation of the valence of the expression based on a scope relationship between at least two terms contained in the expression, wherein the representation stores data representing an attitude of at least a first speaker toward a reported event and toward at least a second speaker; and

associating the representation with the expression.

14. The storage medium of claim 13, further comprising instructions for determining whether at least a first term contained in the expression falls within a scope of at least a second term contained in the expression, wherein the instructions for determine scope include instructions for determine scope by using statistical methods or by analyzing surface punctuation occurring with the expression.

15. The storage medium of claim 14, further comprising instructions for determining whether the second term shifts a valence associated with the first term, and further comprising instructions for marking the representation to incorporate the shifted valence of the first term.

16. The storage medium of claim 15, wherein the instructions for determining whether the second term shifts a valence include instructions for calculating a valence shift based on at least one factive, counter-factive, or implicative appearing in the first term or second term.

17. The storage medium of claim 15, wherein the instructions for determining whether the second term shifts a valence include instructions for calculating a valence shift based on a valence mismatch occurs between the first term and the second term.

18. The storage medium of claim 14, further comprising instructions for determining whether the second term establishes who spoke the first term, and further comprising instructions for marking the representation to indicate who spoke the first term.

19. The storage medium of claim 14, further comprising instructions for determining whether the second term establishes a valence for the first term, and further comprising instructions for marking the representation to indicate the valence of the first term.

20. The storage medium of claim 19, wherein the first term is an unknown term, and wherein the first term inherits valence from the second term, by virtue of a scope relationship between the first and second terms.