RELATING SIMILAR TERMS FOR INFORMATION RETRIEVAL

• US 20090157645A1
• Filed: 12/12/2007
• Published: 06/18/2009
• Est. Priority Date: 12/12/2007
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Alexander, Jeffrey H., Green, Stephen Joseph, Horan, Bernard
• Granted Patent: US 8,001,122 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/48 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually

Abstract

A resource analyzer selects a resource (e.g., document) from a grouping of resources. The grouping of resources can be any type of social tagging system used for information retrieval. The selected resource has an assigned uncontrolled tag and an assigned controlled tag. The controlled tag is a term derived from a controlled vocabulary of terms. Having selected the resource for analyzing, the resource analyzer identifies a first set of resources in the grouping of resources having also been assigned a same value as the uncontrolled tag as the selected resource. Similarly, the resource analyzer identifies a second set of resources in the grouping of resources having also been assigned a same value as the controlled tag. With this information, the resource analyzer then produces a comparison result indicative of a similarity between the first set of resources and the second set of resources.

Claims

1. A method comprising:

selecting a resource from a grouping of resources, the resource having an assigned uncontrolled tag and an assigned controlled tag;

identifying a first set of resources in the grouping of resources having also been assigned a same value as the uncontrolled tag;

identifying a second set of resources in the grouping of resources having also been assigned a same value as the controlled tag; and

producing a comparison result indicative of a similarity between the first set of resources and the second set of resources.

2. The method as in claim 1, wherein the grouping of resources is a grouping of documents, the method further comprising:

calculating a first ranking of terms for the first set of resources, the first ranking indicating a relative importance of distinctive terms contained in each document in the first set of resources; and

calculating a second ranking of terms for the second set of resources, the second ranking indicating a relative importance of distinctive terms contained in each document in the second set of resources.

3. The method as in claim 2 further comprising:

comparing terms in the first ranking with the terms in the second ranking; and

in response to the comparing, determining a similarity value indicative of a relative similarity among the terms in the first ranking and the terms in the second ranking.

4. The method as in claim 3 further comprising:

conditionally generating a rule based on the similarity value and a threshold value, the rule being generated in response to determining that the similarity value exceeds the threshold value, wherein the rule specifies that for any documents that are assigned the uncontrolled tag and do not have an associated controlled tag, those documents will be assigned the controlled tag.

5. The method as in claim 1, wherein producing the comparison result indicative of the similarity between the first set of resources and the second set of resources comprises:

conditionally generating a mapping based on a similarity of attributes in the first set of resources and attributes in the second set of resources, the mapping being generated in response to determining that attributes associated with the first set of resources are sufficiently similar to attributes associated with the second set of resources; and

configuring the mapping to specify that, for a given resource having been assigned the uncontrolled tag, the given resource should also be associated with the controlled tag.

6. The method of claim 5 further comprising:

dynamically applying the mapping to a resource that has already been assigned the uncontrolled tag in the grouping of resources, the application of the mapping associating the controlled tag with the resource that has already been assigned the uncontrolled tag.

7. The method as in claim 1 further comprising:

calculating a first classifier metric for the first set of resources, the first classifier metric indicative of a vector representation of attributes related to the first set of resources that distinguish the first set of resources from the grouping of resources; and

calculating a second classifier metric for the second set of resources, the second classifier metric indicative of a vector representation of attributes related to the second set of resources that distinguish the second set of resources from the grouping of resources.

8. The method as in claim 7 further comprising:

calculating a cosine similarity measurement between the first classifier metric and the second classifier metric, the cosine similarity measurement indicative of a relative similarity between the first set of resources and the second set of resources; and

conditionally generating a rule based on the cosine similarity measurement and a threshold value, the rule being generated in response to determining that the cosine similarity measurement exceeds the threshold value.

9. The method as in claim 7, wherein the grouping of resources is a grouping of documents, the method further comprising:

processing the first classifier metric with respect to the second classifier metric to determine whether the first classifier metric is classifiable in the second set of resources; and

conditionally generating a rule based on processing the first classifier metric against the second classifier metric, the rule being generated in response to determining that the first classifier metric is classifiable in the second set of resources.

10. The method as in claim 2, wherein calculating the first ranking of terms for the first set of resources includes:

for each term in each respective document in the first set of resources, calculating a relative importance value for a given term by performing the steps of;

determining a frequency of occurrence vale for the given term in the document;

determining a group occurrence value indicative of a number of documents in the grouping of resources in which the given term occurs; and

using the frequency of occurrence value and group occurrence value to calculate the relative importance value for the given term.

11. The method as in claim 2, wherein calculating the second ranking of terms for the second set of resources includes:

for each term in each respective document in the second set of resources, calculating a relative importance value for a given term by performing the steps of;

determining a frequency of occurrence value for the given term in the document;

determining a group occurrence value indicative of a number of documents in the grouping of resources in which the given term occurs; and

using the frequency of occurrence value and the group occurrence value to calculate the relative importance value for the given term.

12. The method as in claim 1, wherein the grouping of resources is non-textual media including at least one of an audio resource and an image resource.

13. The method as in claim 1, wherein the controlled tag is a textual term derived from a predefined vocabulary associated with the grouping of resources, and the uncontrolled tag is a textual term not necessarily derived from the predefined vocabulary.

14. The method as in claim 1 further comprising:

in response to identifying the first set of resources in the grouping of resources having also been assigned the same value as the uncontrolled tag, generating the first set of resources to include a first resource and a second resource, each of which has a respective uncontrolled tag of the same value as the uncontrolled tag of the selected resource;

in response to identifying a second set of resources in the grouping of resources having also been assigned the same value as the controlled tag, generating the second set of resources to include the second resource and a third resource, each of which has a respective controlled tag of the same value as the controlled tag of the selected resource.

15. The method as in claim 14, wherein producing the comparison result further comprises:

generating a first attribute list including unique attributes associated with the first set of resources;

generating a second attribute list including unique attributes associated with the second set of resources; and

comparing the first attribute list with the second attribute list to produce the comparison result.

16. A computer system comprising:

a processor;

a memory unit that stores instructions associated with an application executed by the processor; and

an interconnect coupling the processor and the memory unit, enabling the computer system to execute the application and perform operations of;

selecting a resource from a grouping of resources, the resource having an assigned uncontrolled tag and an assigned controlled tag;

identifying a first set of resources in the grouping of resources having also been assigned a same value as the uncontrolled tag;

identifying a second set of resources in the grouping of resources having also been assigned a same value as the controlled tag; and

producing a comparison result indicative of a similarity between the first set of resources and the second set of resources.

17. The computer system as in claim 16, wherein the grouping of resources is a grouping of documents, the computer system further supporting operations of:

calculating a first ranking of terms for the first set of resources, the first ranking indicating a relative importance of distinctive terms contained in each document in the first set of resources;

calculating a second ranking of terms for the second set of resources, the second ranking indicating a relative importance of distinctive terms contained in each document in the second set of resources;

comparing terms in the first ranking with the terms in the second ranking; and

in response to the comparing, determining a similarity value indicative of a relative similarity among the terms in the first ranking and the terms in the second ranking.

18. The computer system as in claim 16, wherein producing the comparison result indicative of the similarity between the first set of resources and the second set of resources comprises:

conditionally generating a rule based on a similarity of attributes in the first set of resources and attributes in the second set of resources, the rule being generated in response to determining that attributes associated with the first set of resources are sufficiently similar to attributes associated with the second set of resources; and

configuring the rule to specify that for any resources that have been assigned a value of the uncontrolled tag and do not have an assigned controlled tag, those resources should also be assigned a value of the controlled tag.

19. A tangible computer-readable medium having instructions stored thereon, the instructions, when executed by a processing device, enabling the processing device to perform operations of:

selecting a resource from a grouping of resources, the resource having an assigned uncontrolled tag and an assigned controlled tag;

identifying a first set of resources in the grouping of resources having also been assigned a same value as the uncontrolled tag;

identifying a second set of resources in the grouping of resources having also been assigned a same value as the controlled tag; and

producing a comparison result indicative of a similarity between the first set of resources and the second set of resources.

20. The tangible computer-readable medium as in claim 19, wherein the grouping of resources is a grouping of documents, the tangible computer-readable medium further comprising:

calculating a first ranking of terms for the first set of resources, the first ranking indicating a relative importance of distinctive terms contained in each document in the first set of resources;

calculating a second ranking of terms for the second set of resources, the second ranking indicating a relative importance of distinctive terms contained in each document in the second set of resources;

comparing terms in the first ranking with the terms in the second ranking;

in response to the comparing, determining a similarity value indicative of a relative similarity among the terms in the first ranking and the terms in the second ranking; and

wherein producing the comparison result indicative of the similarity between the first set of resources and the second set of resources comprises;

conditionally generating a rule based on a similarity of attributes in the first set of resources and attributes in the second set of resources, the rule being generated in response to determining that attributes associated with the first set of resources are sufficiently similar to attributes associated with the second set of resources; and

configuring the rule to specify that for any resources that have been assigned a value of the uncontrolled tag and do not have an assigned controlled tag, those resources should also be assigned a value of the controlled tag.