Composing text and structured databases

• US 8,996,539 B2
• Filed: 07/30/2012
• Issued: 03/31/2015
• Est. Priority Date: 04/13/2012
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Agrawal, Rakesh, Kannan, Anitha, Shafer, John C., Fuxman, Ariel
• Primary Examiner(s): HO, BINH VAN
• Application Number:
  US13/561,085
• Publication Number:
  US 20130275441A1
• Time in Patent Office: 974 Days
• Field of Search: 707/749
• US Class Current: 707/749
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing

Abstract

A framework is provided for composing texts about objects with structured information about these objects, and thus disclosed are methodologies for linking information from at least two data sources—one comprising a plurality of documents comprising text pertaining to at least one object, and one comprising a plurality of structured records comprising at least one characteristic of the at least one object, each characteristic comprising one property name and an associated property value corresponding to the property name for the at least one object—by determining one or more instance-based traits for each object in both data sources and associating at least one record with at least one document that refers to each object, each trait comprising one or more characteristics that identifiably distinguish each object from all other objects.

Claims

1. A system for linking information from at least two data sources, the system comprising:

a first data source comprising a plurality of documents comprising text pertaining to at least one object;

a second data source comprising a plurality of structured records comprising at least one characteristic of the at least one object, each characteristic comprising one property name and an associated property value corresponding to the property name for the at least one object;

a processor for determining one or more traits for each object and for associating at least one record in the second data source with the at least one document from the first data source that refers to each object, wherein each trait is instance-based and comprises at least one characteristic that serves as a proxy for identifying each object from all other objects in the plurality of documents, and wherein at least one of the one or more traits has a different number of characteristics than another trait; and

wherein the system determines that an association of a first trait to a first record is correct if a first text from a first document pertains to either the first record or to an accessory of the product represented by the first record.

2. The system of claim 1, wherein each trait that identifiably distinguishes the at least one object from among all other objects comprises a combination of two or more characteristics that serve as the proxy.

3. The system of claim 1, further comprising a mapping function, performed by the processor, to associate traits in the structured records and the plurality of documents.

4. The system of claim 3, further comprising a scoring function, performed by the processor, to score at least one text document'"'"'s relevancy to at least one record that shares at least one trait.

5. The system of claim 4, wherein the scoring performed by the processor utilizes at least one additional characteristic present in the plurality of documents and in the structured records but not included in the traits.

6. The system of claim 1, wherein the processor further computes an N-gram computation derived from a corpus in order to identify at least one invalid trait, and wherein the processor further prunes at least one instance-based trait based on the N-gram computation.

7. The system of claim 6, wherein the processor further determines at least one instance-based trait for the at least one object that would enable at least one record in the second data source to be associated with the at least one document from the first data source that refers to the at least one object.

8. The system of claim 1, wherein the second data source comprises a plurality of web-based pages.

9. The system of claim 8, wherein a set of objects from the second data source pertain to products.

10. A method for linking information from at least two data sources, the method comprising:

describing a plurality of objects using a set of characteristics as a plurality of traits, wherein each trait in the plurality of traits is instance-based and serves as a proxy to identify each object from other objects in the plurality of objects, and wherein at least one of the plurality of traits has a different number of characteristics than another trait;

using at least one additional characteristic that does not comprise the set of characteristics to score a relevancy of a document to a record from a first data source that shares at least one of the characteristics of another document from a second data source, wherein the additional characteristic is from the document; and

determining that an association of a first trait to a first record is correct if a first text from a first document pertains to either the first record or to an accessory of the product represented by the first record.

11. The method of claim 10, wherein a trait from among the plurality of traits comprises at least one anchor property, and wherein the at least one anchor property is determined by an anchor score corresponding to a relative distinctiveness of a value for at least one corresponding property name in the trait.

12. The method of claim 10, wherein at least one trait from among the plurality of traits is filtered out using an N-Gram analysis, wherein each trait comprises a characteristic from the set of characteristics.

13. The method of claim 10, further comprising precluding at least one characteristic that is ambiguous in view of the first data source.

14. The method of claim 10, further comprising composing information pertaining to a set of objects common to the two data sources based on set the characteristics.

15. A method comprising:

mapping a plurality of documents in a first data source to a plurality of traits in a second data source, wherein each trait in the plurality of traits comprises one or more characteristics, is instance-based, and serves comprises as a proxy for uniquely identifying each object from among a plurality of objects, and wherein at least one of the plurality of traits has a different number of characteristics than another trait;

scoring the relevancy of a document from among the plurality of documents to at least one record from the second data source; and

determining that an association of a first trait to a first record is correct if a first text from a first document pertains to either the first record or to an accessory of the product represented by the first record.

16. The method of claim 15, further comprising identifying the plurality of traits from among the traits present in both the first data source and the second data source.

17. The method of claim 15, further comprising composing the first data source and the second data source based on the mapping and scoring.

18. The method of claim 15, wherein each trait from among the plurality of traits comprises at least one anchor property that operates as a pseudo-key in the second data source.

19. The method of claim 18, wherein the at least one anchor property is determined by an anchor score corresponding to a relative distinctiveness of a value for at least one corresponding property name in the trait.