Automated determination of document utility for a document corpus

• US 10,372,714 B2
• Filed: 02/05/2016
• Issued: 08/06/2019
• Est. Priority Date: 02/05/2016
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Beller, Charles Evan, Dubyak, William G, Sakthi, Palani, Summers, Kristen Maria
• Primary Examiner(s): Ly, Cheyne D
• Application Number:
  US15/017,089
• Publication Number:
  US 20170228434A1
• Time in Patent Office: 1,278 Days
• Field of Search: 707739, 704 7
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/332 Query formulation
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 40/279 Recognition of textual entities

Abstract

A candidate document is received, for example, by a document filter. A determination is made based on the content of the candidate document, whether the candidate document is relevant to a document corpus. A determination is made based on the content of the candidate document, whether the candidate document is novel with respect to the document corpus. In response to determining that the candidate document is relevant to the document corpus and novel with respect to the document corpus, the candidate document is added to the document corpus to make at least a portion of the content of the candidate document available for a response to a search query.

Claims

1. A computer implemented method for determining whether to add a document to an electronic document corpus maintained in at least one electronic data store, the method comprising:

creating, based at least in part on the content of the document corpus, a corpus vector including corpus cells, each corpus cell indicating a word in the document corpus and corpus word count indicating a number of instances of the word in the document corpus; and

receiving, via one or more processors, an electronic candidate document;

creating, based at least in part on the content of the candidate document, a document vector including document cells, each document cell indicating a word in the candidate document and document word count indicating a number of instances of the word in the candidate document;

determining, by at least one of the processors, a relevance value indicating relevance of the candidate document to the document corpus, wherein determining the relevance value includes

for each corpus cell in the document corpus vector, determining a product by multiplying the corpus word count by a document word count of a corresponding document cell in the document vector;

determining a numerator by summing the products;

determining a first sum by summing a square of the corpus word count in each corpus cell in the corpus vector,

determining a second sum by summing a square of the document word count in each document cell in the document vector,

determining a first square root of first sum and a second square root of the second sum;

determining a denominator by multiplying the first square root by the second square root;

dividing the numerator by the denominator;

determining, by the one or more processors, that the candidate document is novel with respect to the document corpus based on the relevance value; and

in response to determining that the candidate document is relevant to the document corpus and novel with respect to the document corpus, adding the candidate document to the document corpus to make at least a portion of the content of the candidate document available for a response to a search query.

2. The method of claim 1, wherein determining, based at least in part on the content of the candidate document, that the candidate document is novel with respect to the document corpus includes determining that the candidate document is novel with respect to one or more documents in the document corpus.

3. The method of claim 2, further comprising randomly selecting the one or more documents from the document corpus.

4. The method of claim 2, wherein creating the candidate document vector includes:

determining first vector cells in the candidate document vector associated with unique words in the candidate document, the first vector cells determined, based at least in part, on counts of the unique words in the candidate document; and

adding a second vector cell to the candidate document vector, the second vector cell comprising a count of unique words not found in the document corpus.

5. The method of claim 4, wherein the first vector cells comprise the counts of the unique words.

6. The method of claim 5, further comprising weighting the counts in the first vector cells according to a term frequency-inverse document frequency statistic.

7. A computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to determine document utility, the program instructions comprising:

program instructions to create, based at least in part on the content of the document corpus, a corpus vector including corpus cells, each corpus cell indicating a word in the document corpus and corpus word count indicating a number of instances of the word in the document corpus; and

program instructions to receive, via one or more processors, an electronic candidate document;

program instructions to create, based at least in part on the content of the candidate document, a document vector including document cells, each document cell indicating a word in the candidate document and document word count indicating a number of instances of the word in the candidate document;

program instructions to determine, by at least one of the processors, a relevance value indicating relevance of the candidate document to a the document corpus, wherein the program instructions to determine the relevance value includes program instructions to

for each corpus cell in the document corpus vector, determining a product by multiply the corpus word count by a document word count of a corresponding candidate document cell;

determine a numerator by summing the products;

determine a first sum by summing a square of the corpus word count in each corpus cell in the corpus vector,

determine a second sum by summing a square of the document word count in each document cell in the document vector,

determine a first square root of first sum and a second square root of the second sum;

determine a denominator by multiplying the first square root by the second square root;

divide the numerator by the denominator;

program instructions to determine, by the one or more processors, that the candidate document is novel with respect to the document corpus based on the relevance value; and

program code to in response to the determination that the candidate document is relevant to the document corpus and novel with respect to the document corpus, add the candidate document to the document corpus to make at least a portion of the content of the candidate document available for a response to a search query.

8. The computer program product of claim 7, the program instructions further comprising program instructions to randomly select the one or more documents from the document corpus.

9. The computer program product of claim 7, wherein the program instructions to create the candidate document vector include program instructions to:

determine first vector cells in the candidate document vector associated with unique words in the candidate document, the first vector cells determined, based at least in part, on counts of the unique words in the candidate document; and

add a second vector cell to the candidate document vector, the second vector cell comprising a count of unique words not found in the document corpus.

10. The computer program product of claim 9, wherein the first vector cells comprise the counts of the unique words.

11. An apparatus comprising:

one or more processors;

a computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by the one or more processors to cause the computer to determine document utility, the program instructions comprising;

program instructions to create, based at least in part on the content of the document corpus, a corpus vector including corpus cells, each corpus cell indicating a word in the document corpus and corpus word count indicating a number of instances of the word in the document corpus; and

program instructions to receive, via one or more processors, an electronic candidate document;

program instructions to create, based at least in part on the content of the candidate document, a document vector including document cells, each document cell indicating a word in the candidate document and document word count indicating a number of instances of the word in the candidate document;

program instructions to determine, by at least one of the processors, a relevance value indicating relevance of the candidate document to a the document corpus, wherein the program instructions to determine the relevance value includes program instructions to

for each corpus cell in the document corpus vector, determining a product by multiply the corpus word count by a document word count of a corresponding document cell;

determine a numerator by summing the products;

determine a first sum by summing a square of the corpus word count in each corpus cell in the corpus vector,

determine a second sum by summing a square of the document word count in each document cell in the document vector,

determine a first square root of first sum and a second square root of the second sum;

determine a denominator by multiplying the first square root by the second square root;

divide the numerator by the denominator;

program instructions to determine, by the one or more processors, that the candidate document is novel with respect to the document corpus based on the relevance value; and

program code to in response to the determination that the candidate document is relevant to the document corpus and novel with respect to the document corpus, add the candidate document to the document corpus to make at least a portion of the content of the candidate document available for a response to a search query.

12. The apparatus of claim 11, the program instructions further comprising program instructions to randomly select the one or more documents from the document corpus.

13. The apparatus of claim 11, wherein the program instructions to create the candidate document vector include program instructions to:

determine first vector cells in the candidate document vector associated with unique words in the candidate document, the first vector cells determined, based at least in part, on counts of the unique words in the candidate document; and

add a second vector cell to the candidate document vector, the second vector cell comprising a count of unique words not found in the document corpus.

14. The apparatus of claim 13, wherein the first vector cells comprise the counts of the unique words.