System and method for providing classification suggestions using document injection

• US 8,909,647 B2
• Filed: 08/19/2013
• Issued: 12/09/2014
• Est. Priority Date: 07/28/2009
• Status: Active Grant

• Current Assignee: Nuix North America Inc. (Nuix Ltd.)
• Original Assignee: FTI Consulting Incorporated
• Inventors: Knight, William C.
• Primary Examiner(s): Wong, Leslie
• Application Number:
  US13/970,560
• Publication Number:
  US 20130339360A1
• Time in Patent Office: 477 Days
• Field of Search: 707/737, 707/740, 707/E17.09, 707/999.006, 707/999.107
• US Class Current: 707/737
• CPC Class Codes: G06F 16/287 Visualization; Browsing
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 16/358 Browsing; Visualisation therefor
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 16/954 Navigation, e.g. using categorised browsing
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/047 Pattern matching networks; Rete networks
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

A system and method for providing classification suggestions using document injection is provided. Clusters of uncoded documents are accessed. A set of reference documents is obtained. Each reference document is associated with a classification code. A set of the uncoded documents selected from one or more of the clusters is identified and compared with the set of reference documents. Those reference documents that are similar to the set of uncoded documents are identified and injected into one or more of the clusters from which the set of uncoded documents is selected. The clusters and a visual suggestion for classification of at least one of the uncoded documents within one of the clusters are displayed.

Claims

1. A system for providing classification suggestions using document injection, comprising:

clusters each comprising uncoded documents;

a set of reference documents, each reference document associated with a classification code;

a set of the uncoded documents selected from one or more of the clusters;

a comparison module to compare the set of uncoded documents with the set of reference documents;

an identification module to identify those reference documents that are similar to the set of uncoded documents;

an injection module to inject the similar reference documents into one or more of the clusters from which the set of uncoded documents are selected;

a display to display the clusters and to provide a visual suggestion for classification of at least one of the uncoded documents within one of the clusters based on the similar reference documents in that cluster; and

a classification assignment module to count a number of reference documents within the cluster for each different type of classification code, to determine a distance between the uncoded document and each of the reference documents in the cluster, to weigh the count of the reference documents for each type of classification code based on the distances of the reference documents associated with that classification code, and to assign the classification code having the highest weighted count to the uncoded document.

2. The system according to claim 1, further comprising:

a spine module to organize the clusters along one or more vectors as cluster spines.

3. The system according to claim 2, further comprising:

a similarity module to determine the similar reference documents based on the cluster spines to generate a score vector for at least one of the cluster spines, to compare the spine score vector with score vectors associated with the reference documents, and to identify the similar reference documents based on a similarity of the spine score vector with the score vectors for the reference documents.

4. The system according to claim 1, further comprising:

a compass placed within the cluster display; and

a visual module to visually emphasize those clusters located within the compass and to visually deemphasize those clusters appearing outside of the compass.

5. The system according to claim 4, further comprising:

labels placed circumferentially around the compass, wherein each label corresponds to one or more concepts that describe a grouping of one or more of the clusters appearing within the compass.

6. The system according to claim 1, further comprising:

a document presentation module to display each document as a circle within the display and to display the classification suggestion of the uncoded document by filling at least a portion of the circle with one or more of a color and a shape designating that classification suggestion.

7. The system according to claim 1, further comprising:

a classification assignment module to select one other of the uncoded documents within one of the clusters, to identify one of the reference documents located closest to the other uncoded document and to assign the classification code of the closest reference document to the other uncoded document as the classification suggestion.

8. The system according to claim 1, further comprising:

a classification assignment module to select one other of the uncoded documents within one of the clusters, to count a number of the reference documents within the cluster for each different type of classification code and to provide to the other uncoded document the classification code having the highest count of reference documents as the suggestion for classification.

9. The system according to claim 1, further comprising:

a classification assignment module to select one other of the uncoded documents within one of the clusters, to calculate an average distance of the reference documents within the cluster for each different type of classification code and to assign to the other uncoded document the classification code with the closest average distance to that other uncoded document.

10. The system according to claim 1, further comprising:

a classification assignment module to select one other of the uncoded documents within one of the clusters, to count a number of reference documents within the cluster for each different type of classification code, to determine a distance between the other uncoded document and each of the reference documents in the cluster, to weigh the count of the reference documents for each type of classification code based on the distances of the reference documents associated with that classification code, and to assign the classification code having the highest weighted count to the other uncoded document.

11. A method for providing classification suggestions using document injection, comprising:

accessing clusters each comprising uncoded documents;

obtaining a set of reference documents, each reference document associated with a classification code;

identifying a set of the uncoded documents selected from one or more of the clusters;

comparing the set of uncoded documents with the set of reference documents;

identifying those reference documents that are similar to the set of uncoded documents;

injecting the similar reference documents into one or more of the clusters from which the set of uncoded documents are selected;

displaying the clusters and providing a visual suggestion for classification of at least one of the uncoded documents within one of the clusters based on the similar reference documents in that cluster; and

for each different type of classification code, counting a number of reference documents within the cluster for that classification code;

determining a distance between the uncoded document and each of the reference documents in the cluster;

weighing the count of the reference documents for each type of classification code based on the distances of the reference documents associated with that classification code; and

assigning the classification code having the highest weighted count to the uncoded document.

12. The method according to claim 11, further comprising:

organizing the clusters along one or more vectors as cluster spines.

13. The method according to claim 12, further comprising:

determining the similar reference documents based on the cluster spines, comprising;

generating a score vector for at least one of the cluster spines;

comparing the spine score vector with score vectors associated with the reference documents; and

identifying the similar reference documents based on a similarity of the spine score vector with the score vectors for the reference documents.

14. The method according to claim 11, further comprising:

providing a compass within the cluster display;

visually emphasizing those clusters located within the compass; and

visually deemphasizing those clusters appearing outside of the compass.

15. The method according to claim 14, further comprising:

providing labels circumferentially around the compass, wherein each label corresponds to one or more concepts that describe a grouping of one or more of the clusters appearing within the compass.

16. The method according to claim 11, further comprising:

displaying each document as a circle within the display; and

displaying the classification suggestion of the uncoded document by filling at least a portion of the circle with one or more of a color and a shape designating that classification suggestion.

17. The method according to claim 11, further comprising:

selecting one other of the uncoded documents within one of the clusters;

identifying one of the reference documents located closest to the other uncoded document; and

assigning the classification code of the closest reference document to the other uncoded document as the classification suggestion.

18. The method according to claim 11, further comprising:

selecting one other of the uncoded documents within one of the clusters;

for each different type of classification code, counting a number of the reference documents associated with that classification code in the cluster; and

providing to the other uncoded document, the classification code having the highest count of reference documents as the suggestion for classification.

19. The method according to claim 11, further comprising:

selecting one other of the uncoded documents within one of the clusters;

for each different type of classification code, calculating an average distance of the reference documents associated with that classification code in the cluster; and

assigning to the other uncoded document the classification code with the closest average distance to that other uncoded document.

20. The method according to claim 11, further comprising:

selecting one other of the uncoded documents within one of the clusters;

for each different type of classification code, counting a number of reference documents within the cluster for that classification code;

determining a distance between the other uncoded document and each of the reference documents in the cluster;

weighing the count of the reference documents for each type of classification code based on the distances of the reference documents associated with that classification code; and

assigning the classification code having the highest weighted count to the other uncoded document.