Information retrieval system for archiving multiple document versions

• US 8,612,427 B2
• Filed: 03/04/2010
• Issued: 12/17/2013
• Est. Priority Date: 01/25/2005
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Patterson, Anna L.
• Primary Examiner(s): Vy, Hung T
• Application Number:
  US12/717,663
• Publication Number:
  US 20100169305A1
• Time in Patent Office: 1,384 Days
• Field of Search: 707/705, 707/709, 707/722, 707723-728, 707/736, 707/758, 707/661, 707/667, 715/500, 715/203, 709/224
• US Class Current: 707/725
• CPC Class Codes: G06F 16/219 Managing data history or versioning querying versioned data G06F16/2474; querying temporal data G06F16/2477
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 40/197 Version control for software G06F8/71
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

An information retrieval system uses phrases to index, retrieve, organize and describe documents. Phrases are identified that predict the presence of other phrases in documents. Documents are the indexed according to their included phrases. Index data for multiple versions or instances of documents is also maintained. Each document instance is associated with a date range and relevance data derived from the document for the date range.

Claims

1. A method, performed by at least one computer system, of retrieving documents in response to a search query submitted on a first date that includes one or more phrases, the method comprising:

selecting from a corpus of documents a plurality of documents that include at least one of the one or more phrases and that are relevant to the query based on a relevance score for the selected documents;

determining, using a processor of the at least one computer system, for at least one of the selected documents, a date range for the document, the date range representing a period for which no change in the document has been detected;

calculating, using a processor of the at least one computer system, a weighted relevance score for each of the selected documents, the weighted relevance score for a particular document being based on the relevance score for the particular document and a difference between the first date and the date range for the particular document; and

ranking, using a processor of the at least one computer system, the selected documents according to their weighted relevance scores.

2. The method of claim 1, wherein the relevance score for each of the selected documents is based on phrases contained in the document during the date range.

3. The method of claim 1, wherein the relevance score for each of the selected documents is based on one or more from the group consisting of:

document length of the document during the date range, a frequency of a phrase contained in the document during the date range, types of HTML markup in the document during the date range.

4. The method of claim 1, wherein the relevance score is up-weighted for older documents.

5. The method of claim 1, further comprising:

determining that relevance scores for documents with a date range including a historical date are higher than relevance scores for documents having a date range including the first date; and

down-weighting the relevance scores for documents having a date range including the first date as part of calculating the weighted relevance score for each of the selected documents.

6. The method of claim 1, wherein the search query includes a date term and the weighted relevance score for the particular document is based on a function of the date term and the date range for the particular document.

7. The method of claim 1, wherein the relevance score is down-weighted for older documents.

8. The method performed of claim 1, further comprising:

determining, using a processor of the at least one computer system, a date term included in the search query;

wherein each selected document has an instance with a date range that includes the date term.

9. A method performed by at least one computer system, of retrieving documents in response to a search query that includes one or more phrases, the method comprising:

selecting from a corpus of documents a plurality of documents that include at least one of the one or more phrases and that are relevant to the query based on relevance scores for the selected documents;

determining, using a processor of the at least one computer system, for each selected document a date range with which the document is associated, the date range representing a period for which no change in the document has been detected;

grouping the documents into a plurality of groups using to the date range associated with each document; and

within each group, ranking the documents according to relevance scores determined based on relevance data valid for the date range associated with each document.

10. The method of claim 9, wherein each group represents a month, newer groups are presented ahead of older groups, and within each group, higher relevance scores are presented ahead of lower relevance scores.

11. A method, performed by at least one computer system, of retrieving documents in response to a search query that includes one or more phrases, the method comprising:

accessing an index for a corpus of documents, the corpus comprising documents having multiple instances, each instance of a particular document being acquired at a different date and a new instance being acquired when a change from a previous instance is detected, wherein each instance has an associated date range;

selecting from the index a plurality of documents that are relevant to the search query;

determining, using a processor of the at least one computer system, for each instance of a selected document, a relevance score based on relevance data valid for the instance of the selected document and its associated date range;

selecting, for each document, the document instance with the highest relevance score; and

ranking the selected document instances according to their relevance scores.

12. The method of claim 11, wherein each instance of the particular document may be selected independently of another instance.

13. The method of claim 11, wherein each instance of the document has an identifier and the index stores information for each instance using its respective identifier.

14. The method of claim 13, wherein the document identifier includes a date stamp.

15. The method of claim 14, wherein the date stamp is a hash of the date the instance was indexed.

16. The method of claim 11, wherein determining the relevance score for each instance of a selected document includes:

determining that relevance scores for instances of documents with a date range including a historical date are higher than relevance scores for instances of documents with a date range including the date the query was submitted; and

down-weighting the relevance scores for instances having a date range including the date the query was submitted.

17. A computer system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the computer system to perform operations comprising;

selecting from a corpus of documents a plurality of documents that include at least one phrase of a search query received by the computer system and that are relevant to the search query, the selection being based on a relevance score for the selected documents;

determining for at least one of the selected documents a date range for the document, the date range representing a period for which no change in the document has been detected;

calculating a weighted relevance score for each of the selected documents, the weighted relevance score for a particular document being based on the relevance score for the particular document and a difference between a date the query was submitted and the date range for the particular document; and

ranking the selected documents according to their weighted relevance scores.

18. The system of claim 17, wherein the relevance score is based on one or more from the group consisting of:

document length of the document during the date range, a frequency of a phrase contained in the document during the date range, types of HTML markup in the document during the date range.

19. The system of claim 17, the operations further comprising:

determining that relevance scores for documents with a date range including a historical date are higher than relevance scores for documents with a date range including the date the query was submitted; and

down-weighting the relevance scores for documents with a date range including the date the query was submitted as part of calculating the weighted relevance score for each of the selected documents.

20. The system of claim 17, wherein the date range includes a first value indicating the date the document was indexed and a second value, the second value being a token indicating a currently valid document when the date range includes the date the query was submitted.

21. A computer system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the computer system to perform operations comprising;

selecting from a corpus of documents a plurality of documents that include at least one phrase of a search query and that are relevant to the query based on relevance scores for the selected documents;

determine for each selected document a date range with which the document is associated, the date range representing a period for which no change in the document has been detected;

grouping the documents into a plurality of groups using the date range associated with each document; and

within each group, ranking the documents according to relevance scores determined based on relevance data valid for the date range associated with each document.

22. A computer system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the computer system to perform operations comprising;

accessing an index for a corpus of documents, the corpus comprising documents having multiple instances, each instance of a particular document being acquired at a different date and a new instance being acquired when a change from a previous instance is detected, wherein each instance has an associated date range;

selecting from the index a plurality of documents that are relevant to a search query;

determining for each instance of a selected document, a relevance score based on relevance data valid for the instance of the selected document and its associated date range;

selecting for each document the document instance with the highest relevance score; and

ranking the selected document instances according to their relevance scores.

23. A computer-readable storage device having embodied thereon instructions that, when executed by one or more processors, cause at least one computer system to perform operations comprising:

selecting, from a corpus of documents, a plurality of documents that are responsive to a search query received by the computer system, the selection being based on a relevance score for the selected documents;

determining for at least one of the selected documents a date range for the document, the date range representing a period for which no change in the document has been detected;

calculating a weighted relevance score for each of the selected documents, the weighted relevance score for a particular document being based on the relevance score for the particular document and a difference between a date the query was submitted and the date range for the particular document; and

ranking the selected documents according to their weighted relevance scores.