Matching Offers to Known Products

• US 20110289026A1
• Filed: 05/20/2010
• Published: 11/24/2011
• Est. Priority Date: 05/20/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Givoni, Inmar-Ella, Kannan, Anitha
• Granted Patent: US 8,417,651 B2
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06F 16/38 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/907 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06Q 30/00 Commerce

Abstract

A method and apparatus for electronically matching an electronic offer to structured data for a product offering is disclosed. The structure data is reviewed and a dictionary of terms for each attribute from the structure data is created. Attributes in unstructured text may be determined. Each pair of the attributes (name and value) from the unstructured data and the structured data are obtained, the attribute pairs of the structured data and the unstructured data and compared and a similarity level is calculated for the matching the attribute pairs. The structured data pair that has the highest similarity score to the unstructured data pair is selected and returned.

Claims

1. A computer based method of electronically matching an electronic offer to structured data comprising a product offering comprising:

reviewing the structured data;

creating a dictionary of terms for each attribute from the structured data;

determining attributes in unstructured text comprising

submitting the unstructured text to a parsing system wherein the parsing system parses the unstructured text comprising;

reviewing the text of the unstructured data;

collecting statistical data regarding attribute frequency in the unstructured data;

selecting key attributes from the unstructured text comprising

determining the probability an attribute is a key attribute by studying structured data;

if the probability is above a threshold, marking the attribute as one of the key attributes;

if the probability is below a threshold, not marking the attribute as one of the key attributes;

obtaining attribute level similarity between the unstructured record and structured records for the key attributes comprising;

maximal parsing the unstructured record comprising

for each of the attributes above threshold

 

obtain each pair of the attribute from the unstructured data and the structured data;

 

matching the attribute pairs comprising a name and a value of the structured data and the unstructured data;

 

calculating a similarity level for the matching the attribute pairs comprising a name and a value of the structured data and the unstructured data

 

selecting the structured data pair that has the highest similarity score to the unstructured data pair; and

returning the structured data pair that has the highest similarity score to the unstructured data pair.

2. The method of claim 1, wherein calculating a similarity level further comprising creating a similarity feature vector f length K and populate its elements with similarity levels between the maximum unstructured record and s for the corresponding K (key attribute) based on attribute type to enable the maximum agreement between the attribute pairs of structured data and the attribute pairs of unstructured data.

3. The method of claim 1, wherein if the attribute type is binary, requiring an exact match and if the attribute type is numeric, requiring a best match.

4. The method of claim 1, further comprising training the system with unstructured and structured pairs which are known to be matched and known to be mismatched.

5. The method of claim 1, further comprising segmenting the unstructured records into regions wherein the regions may overlap.

6. The method of claim 1, wherein segmenting the unstructured records into regions wherein the regions may overlap further comprises:

adding tokens to the unstructured text; and

using the dictionary to associate potential attributes to a subset of the tokens.

7. The method of claim 6, wherein the dictionary is directional.

8. The method of claim 1, wherein matches with longer character match runs between unstructured data structured data values and are given more weight than attribute matches with shorter run matches between unstructured data structured data values.

9. A computing system for electronically matching an electronic offer to structured data comprising a product offering comprising a processor physically configured to execute computer executable instructions, a memory for storing the computer executable instructions and an input/output device, the computer executable instructions comprising instructions for:

reviewing the structured data;

creating a dictionary of terms for each attribute from the structured data;

determining attributes in unstructured text comprising

submitting the unstructured text to a parsing system wherein the parsing system parses the unstructured text comprising;

reviewing the text of the unstructured data;

collecting statistical data regarding attribute frequency in the unstructured data;

selecting key attributes from the unstructured text comprising

determining the probability an attribute is a key attribute by studying structured data;

if the probability is above a threshold, marking the attribute as one of the key attributes;

if the probability is below a threshold, not marking the attribute as one of the key attributes;

obtaining attribute level similarity between the unstructured record and structured records for the key attributes comprising;

maximal parsing the unstructured record comprising

for each of the attributes above threshold

 

obtain each pair of the attribute from the unstructured data and the structured data;

 

matching the attribute pairs comprising a name and a value of the structured data and the unstructured data;

 

calculating a similarity level for the matching the attribute pairs comprising a name and a value of the structured data and the unstructured data

 

selecting the structured data pair that has the highest similarity score to the unstructured data pair; and

returning the structured data pair that has the highest similarity score to the unstructured data pair.

10. The computing system of claim 9, wherein the computer executable instructions for calculating a similarity level further comprising computer executable instructions for creating a similarity feature vector f length K and populate its elements with similarity levels between the maximum unstructured record and s for the corresponding K (key attribute) based on attribute type to enable the maximum agreement between the attribute pairs of structured data and the attribute pairs of unstructured data.

11. The computing system of claim 9, wherein if the attribute type is binary, requiring an exact match and if the attribute type is numeric, requiring a best match.

12. The computing system of claim 9, further comprising computer executable instructions for training the system with unstructured and structured pairs which are known to be matched and known to be mismatched.

13. The computing system of claim 9, further comprising computer executable instructions for segmenting the unstructured records into regions wherein the regions may overlap.

14. The computing system of claim 13, wherein segmenting the unstructured records into regions wherein the regions may overlap further comprises:

adding tokens to the unstructured text; and

using the dictionary to associate potential attributes to a subset of the tokens.

15. The computing system of claim 9, wherein the dictionary is directional.

16. The computing system of claim 9, wherein matches with longer character match runs between unstructured data structured data values and are given more weight than attribute matches with shorter run matches between unstructured data structured data values.

17. A computer storage medium physically configured according to computer executable instructions for electronically matching an electronic offer to structured data comprising a product offering, the computer executable instructions comprising instructions for:

reviewing the structured data;

creating a dictionary of terms for each attribute from the structured data;

determining attributes in unstructured text comprising

submitting the unstructured text to a parsing system wherein the parsing system parses the unstructured text comprising;

reviewing the text of the unstructured data;

collecting statistical data regarding attribute frequency in the unstructured data;

selecting key attributes from the unstructured text comprising

determining the probability an attribute is a key attribute by studying structured data;

if the probability is above a threshold, marking the attribute as one of the key attributes;

if the probability is below a threshold, not marking the attribute as one of the key attributes;

obtaining attribute level similarity between the unstructured record and structured records for the key attributes comprising;

maximal parsing the unstructured record comprising

for each of the attributes above threshold

 

obtain each pair of the attribute from the unstructured data and the structured data;

 

matching the attribute pairs comprising a name and a value of the structured data and the unstructured data;

 

calculating a similarity level for the matching the attribute pairs comprising a name and a value of the structured data and the unstructured data

 

selecting the structured data pair that has the highest similarity score to the unstructured data pair; and

returning the structured data pair that has the highest similarity score to the unstructured data pair.

18. The computer storage medium of claim 17, wherein the computer executable instructions for calculating a similarity level further comprising computer executable instructions for creating a similarity feature vector f length K and populate its elements with similarity levels between the maximum unstructured record and s for the corresponding K (key attribute) based on attribute type to enable the maximum agreement between the attribute pairs of structured data and the attribute pairs of unstructured data.

19. The computer storage medium of claim 17, wherein if the attribute type is binary, requiring an exact match and if the attribute type is numeric, requiring a best match.

20. The computer storage medium of claim 17, further comprising computer executable instructions for training the system with unstructured and structured pairs which are known to be matched and known to be mismatched and wherein matches with longer character match runs between unstructured data structured data values and are given more weight than attribute matches with shorter run matches between unstructured data structured data values.