Deriving document similarity indices

• US 8,478,740 B2
• Filed: 12/16/2010
• Issued: 07/02/2013
• Est. Priority Date: 12/16/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Gherman, Sorin, Mukerjee, Kunal, Prout, Adam
• Primary Examiner(s): Coby, Frantz
• Application Number:
  US12/970,650
• Publication Number:
  US 20120158731A1
• Time in Patent Office: 929 Days
• Field of Search: 707/673, 707/696, 707/711, 707/715
• US Class Current: 707/715
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/40 of multimedia data, e.g. slideshows comprising image and additional audio data retrieval of still image data G06F16/50; retrieval of audio data G06F16/60; retrieval of video data G06F16/70
• CPC Class Codes: G06F 16/41 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/93 Document management systems

Abstract

The present invention extends to methods, systems, and computer program products for deriving document similarity indices. Embodiments of the invention include scalable and efficient mechanisms for deriving and updating a document similarity index for a plurality of documents. The number of maintained similarities can be controlled to conserve CPU and storage resources.

Claims

1. At a computer system including one or more processors and system memory, a method for deriving a document similarity index for a plurality of documents, the method comprising:

an act of accessing a document;

an act computing a tag index for the document, the tag index including one more keyword/weight pairs, each keyword/weight pair mapping a keyword to a corresponding weight for the keyword to indicate a significance of the keyword within the document;

an act of identifying a specified number of most significant keywords in the document based on weights in the tag index;

for each keyword in the specified number of the most significant keywords, an act of determining the corresponding weight of the keyword in each document in the plurality of documents;

an act of identifying a plurality of candidate documents, from the among the plurality of documents, based on the corresponding weights of the specified number of the most significant keywords in the plurality of documents, at least some of the specified number of the most significant keywords in the document also being significant keywords in each of the plurality of candidate documents;

for each candidate document in the plurality of candidate documents, an act of calculating a full similarity between the document and candidate document by determining the weight of additional keywords from the document within the candidate document;

an act of selecting full similarities for a prescribed number of a candidate documents for inclusion in the document similarity index to indicate documents that are similar to the document, selection of the full similarities for the prescribed number of candidate documents based on the full similarity calculations and in accordance with one of a hard limit or an express threshold, the hard limit or the express threshold limiting the number of candidate documents that can be selected for inclusion in the document similarity index; and

for each candidate document included in the prescribed number of candidate documents, an act of storing information from the full similarity between the document and the candidate document in the document similarly index.

2. The method as recited in claim 1, wherein the act of accessing a document comprises an act of accessing a document that includes words in a written language.

3. The method as recited in claim 1, wherein an act computing a tag index for the document comprises computing keyword weights based on keyword frequency within the document and document length.

4. The method as recited in claim 1, wherein the act of identifying a plurality of candidate documents, from the among the plurality of documents, based on the corresponding weights of the specified number of the most significant keywords in the plurality of documents comprises an act of accessing at least one keyword/weight pair form a least recently used (“

LRU”

) cache.

5. The method as recited in claim 4, wherein accessing at least one keyword/weight pair from a least recently used (“

LRU”

) cache comprises an act of using an LRU approximation.

6. The method as recited in claim 1, wherein for each candidate document in the plurality of candidate documents, the act of calculating a full similarity between the document and candidate document comprises an act of using a cosine-similarity function to calculate the similarity between the document and the candidate document.

7. The method as recited in claim 1, wherein selecting full similarities for a prescribed number of a candidate documents for inclusion in the document similarity index comprises an act of selecting full similarities for a prescribed number of a candidate documents in accordance with a hard limit that limits the number of similarities that can be selected for inclusion in the document similarity index to ten or less.

8. The method as recited in claim 1, wherein for each candidate document included in the prescribed number of candidate documents, the act of storing information from the full similarity in the document similarly index comprises an act of storing a mapping that quantifies the similarity between the document and the candidate document in a similarity value.

9. At a computer system including one or more processors and system memory, the computer system also including a plurality of documents and a document similarity index, the document similarity index indicating similarities between different documents in the plurality of documents, a method for updating the document similarity index, the method comprising:

an act of accessing a batch of documents;

for each document in the batch of documents, an act of computing a tag index for the document, the tag index including one more keyword/weight pairs, each keyword/weight pair mapping a keyword to a corresponding weight for keyword to indicate a significance of the keyword within the document;

for each document in the batch of documents subsequent to computing the tag indices;

an act of identifying a specified number of the most significant keywords in the document based on weights in the tag index;

for each keyword in the specified number of most significant keywords, an act of determining the corresponding weight of the keyword in each document in the plurality of documents and in the batch of documents;

an act of identifying a plurality of candidate documents, from the among the plurality of documents and the batch of documents, based on the corresponding weights of the specified number of the most significant keywords in the plurality of documents and in the batch of documents, at least some of the specified number of the most significant keywords in the document also being significant keywords in each of the plurality of candidate documents;

for at least one candidate document identified from within the plurality of documents;

an act of calculating a full similarity between the document and candidate document by determining the weight of additional keywords from the document within the candidate document;

an act of identifying the weakest similarity, from among a specified number of top similarities, for the candidate document from within the document similarity index, the weakest similarity indicating the similarity between the candidate document and another document in the plurality of documents;

an act of determining that the candidate document and the document are more similar than the candidate document and the other document by comparing the calculated full similarity to the identified weakest similarity; and

an act of replacing the weakest similarly with information from the calculated full similarity within the document similarity index to incrementally update the document similarity index, the replacement based on the determination; and

for any candidate documents identified from within the batch of documents;

an act of calculating a full similarity between the document and candidate document by determining the weight of additional keywords from the document within the candidate document;

an act of selecting a prescribed number of candidate documents for inclusion in the document similarity index as documents that are similar to the document, selection of the prescribed number of candidate documents based on the full similarity calculations and in accordance with one of a hard limit or an express threshold, the hard limit or the express threshold limiting the number of candidate documents that can be selected for inclusion in the document similarity index; and

an act of storing information from the calculated full similarity between the document and the candidate document in the document similarly index.

10. The method as recited in claim 9, further comprising for at least one other candidate document identified from within the plurality of documents:

an act of calculating a full similarity between the document and candidate document by determining the weight of additional keywords from the document within the candidate document;

an act of identifying the weakest similarity, from among a specified number of top similarities, for the candidate document from within the document similarity index, the weakest similarity indicating the similarity between the candidate document and a second other document in the plurality of documents;

an act of determining that the candidate document and the second other document are more similar than the candidate document and the document by comparing the calculated full similarity to the identified weakest similarity; and

an act of retaining the weakest similarity with in the document similarity index based on the determination.

11. The method as recited in claim 9, wherein the act of accessing a batch of documents comprises an act of accessing a batch of documents that include words in a written language.

12. The method as recited in claim 9, wherein for each document in the batch of documents, the act of computing a tag index for the document comprises an act of computing keyword weights based on keyword frequency within the document and document length.

13. The method as recited in claim 9, wherein the act of replacing the weakest similarly with information from the calculated full similarity within the document similarity index comprises an act of overwriting a similarity value that quantifies the similarity between the candidate document and the other document with a similarity value that quantifies the similarity between the candidate document and the document.

14. The method as recited in claim 9, wherein for any candidate documents identified from within the batch of documents, the act of calculating a full similarity between the document and candidate document comprises an act of using a cosine-similarity function to calculate the similarity between the document and the candidate document.

15. The method as recited in claim 14, wherein selecting full similarities for a prescribed number of a candidate documents for inclusion in the document similarity index comprises an act of selecting full similarities for a prescribed number of a candidate documents in accordance with a hard limit that limits the number of similarities that can be selected for inclusion in the document similarity index to ten or less.

16. The method as recited in claim 15, wherein the act of storing information from the full similarity in the document similarly index comprises an act of storing a mapping that quantifies the similarity between the document and the candidate document in a similarity value.

17. The method as recited in claim 16, wherein for at least one candidate document identified from within the plurality of documents, the act determining that the candidate document and the document are more similar than the candidate document and the other document by comparing the calculated full similarity to the identified weakest similarity comprises an act of comparing a first similarity value quantifying the similarity candidate document and the document to a second similarity value quantifying the similarity candidate document and the other document.

18. A computer program product for use at a computer system, the computer program product a method for deriving a document similarity index for a plurality of documents, the computer program product comprising one or more computer storage devices having stored thereon computer executable instructions that when executed at a processor cause, the computer system to perform the method including the following:

access a document containing words in a written language;

compute a tag index for the document, the tag index including one more keyword/weight pairs, each keyword/weight pair mapping a keyword to a corresponding weight for the keyword to indicate a significance of the keyword within the document;

identify a specified number of most significant keywords in the document based on weights in the tag index;

for each keyword in the specified number of the most significant keywords, determine the corresponding weight of the keyword in each document in the plurality of documents;

identify a plurality of candidate documents, from the among the plurality of documents, based on the corresponding weights of the specified number of the most significant keywords in the plurality of documents, at least some of the specified number of the most significant keywords accessed from a least recently used cache;

for each candidate document in the plurality of candidate documents, an act of use a cosine-similarity function to calculating a full similarity between the document and candidate document by determining the weight of additional keywords from the document within the candidate document;

for a first one or more candidate documents;

select full similarities for a prescribed number of a candidate documents for inclusion in the document similarity index to indicate documents that are similar to the document, selection of the full similarities for the prescribed number of candidate documents based on the full similarity calculations and in accordance with one of a hard limit or an express threshold, the hard limit or the express threshold limiting the number of candidate documents that can be selected for inclusion in the document similarity index; and

for each candidate document included in the prescribed number of candidate documents, store information from the full similarity between the document and the candidate document in the document similarly index,

for a second one or more candidate documents;

identifying the weakest similarity, from among a specified number of top similarities, for the candidate document from within the document similarity index, the weakest similarity indicating the similarity between the candidate document and another document in the plurality of documents;

determine if the candidate document and the document are more similar than the candidate document and the other document by comparing the calculated full similarity to the identified weakest similarity;

replace the weakest similarly with information from the calculated full similarity within the document similarity index, when candidate document and the document are more similar than the candidate document and the other document; and

retain the weakest similarly, when candidate document and the document are not more similar than the candidate document and the other document.

19. The computer program product of claim 18, wherein computer executable instructions that, when executed, cause the computer system to replace the weakest similarly with information from the calculated full similarity within the document similarity index computer executable instructions that, when executed, cause the computer system to overwrite a similarity value in the document similarity index.

20. The computer program product of claim 18, wherein computer executable instructions that, when executed, cause the computer system to retain the weakest similarly comprise computer executable instructions that, when executed, cause the computer system to retain an existing similarity value in the document similarity index.