Suggesting long-tail tags

• US 7,996,418 B2
• Filed: 04/30/2008
• Issued: 08/09/2011
• Est. Priority Date: 04/30/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Weinstein, Alex David, Ryabkov, Dmitry Yevgenyevich
• Primary Examiner(s): Cottingham; John R.
• Assistant Examiner(s): Bromell; Alexandria Y
• Application Number:
  US12/113,187
• Publication Number:
  US 20090276437A1
• Time in Patent Office: 1,196 Days
• Field of Search: 707/767, 707/673
• US Class Current: 707/767
• CPC Class Codes: G06F 16/48 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually

Abstract

Technologies are described herein for suggesting long-tail tags. A first group of tags and a second group of tags are identified from a plurality of tags. The first group of tags includes frequently-assigned tags having a higher frequency of being assigned to an asset. The second group of tags includes long-tail tags having a lower frequency of being assigned to the asset than the frequently-assigned tags. The frequently-assigned tags and a sample of the long-tail tags are suggested to a user upon receiving a request from the user to tag the asset.

Claims

1. A computer-implemented method for suggesting a long-tail tag, the method comprising computer-implemented operations for:

identifying, by a computer having a processor and a memory, a first group of tags from a plurality of tags, the first group of tags comprising frequently-assigned tags having a higher frequency of being assigned to an asset;

identifying, by the computer, a second a group of tags from the plurality of tags, the second group of tags comprising long-tail tags having a lower frequency of being assigned to the asset than the frequently-assigned tags in the first group of tags;

suggesting, by the computer, the frequently-assigned tags and a sample of the long-tail tags to a user upon receiving a request from the user to tag the asset;

identifying, by the computer, a third group of tags from the plurality of tags, the third group of tags comprising rarely-assigned tags having a lower frequency of being assigned to the asset than the long-tail tags in the second group of tags,

tracking, by the computer and over a given number of chances, a number of times that a tag in the second group has been assigned to the asset when the tag is suggested;

determining, by the computer, whether the number of times that the tag has been suggested to the user surpasses a first threshold;

wherein the first threshold comprises a minimum number of times for the tag to be assigned to the asset in order for the tag to be considered one of the frequently-assigned tags,

upon determining that the number of times that the tag has been suggested to the user surpasses the first threshold, removing, by the computer, the tag from the second group of tags and adding the tag into the first group of tags;

determining whether the number of times that the tag has been suggested to the user falls below a second threshold;

wherein the second threshold comprises a maximum number of times for the tag to be assigned to the asset in order for the tag to be considered one of the rarely-assigned tags, and

upon determining that the number of times that the tag has been suggested to the user falls below the second threshold, removing the tag from the second group of tags and adding the tag into the third group of tags.

2. The computer-implemented method of claim 1, wherein each of the plurality of tags comprises one or more keywords.

3. The computer-implemented method of claim 1, wherein identifying a first group of tags from a plurality of tags comprises identifying a first percentage of tags that have a highest frequency of being assigned to the asset, wherein identifying a third group of tags from the plurality of tags comprises identifying a second percentage of tags that have a lowest frequency of being assigned to the asset, and wherein identifying a second a group of tags from the plurality of tags comprises identifying the tags remaining between the first percentage of tags and the second percentage of tags.

4. The computer-implemented method of claim 1, wherein suggesting the frequently-assigned tags and a sample of the long-tail tags to a user upon receiving a request from the user to tag the asset comprises:

selecting, by the computer, a given number of most frequently-assigned tags; and

randomly or pseudo-randomly selecting, by the computer, the sample of the long-tail tags.

5. The computer-implemented method of claim 1, wherein suggesting the frequently-assigned tags and a sample of the long-tail tags to a user upon receiving a request from the user to tag the asset comprises providing, by the computer, the frequently-assigned tags and the sample of the long-tail tags via a web server application over a network for display by a remote computer.

6. A computer-based system for suggesting a long-tail tag, comprising:

a computer having a processor and a memory;

a web server application executed by the computer and adapted to communicate with a remote computer over a network;

a database coupled to the computer and adapted to store a plurality of tags and a number of times each of the plurality of tags has been assigned to an asset;

a tag suggestion module executed by the computer and adapted to suggest, via the web server application, frequently-assigned tags having a higher frequency of being assigned to the asset and a sample of long-tail tags having a lower frequency of being assigned to the asset than the frequently-assigned tags;

wherein the tag suggestion module is further adapted to divide the plurality of tags into a first group containing the frequently-assigned tags having a higher frequency of being assigned to the asset, a second group containing the long-tail tags having a lower frequency of being assigned to the asset than the frequently-assigned tags in the first group, and a third group containing rarely-assigned tags having a lower frequency of being assigned to the asset than the long-tail tags in the second group,

wherein the tag suggestion module is further adapted to track, over a given number of chances, a number of times that a tag in the second group has been assigned to the asset when the tag is suggested;

wherein the tag suggestion module is further adapted to remove the tag from the second group and to add the tag to the first group upon determining that a number of times that the tag has been assigned to the asset exceeds a first threshold, and wherein the tag suggestion module is further adapted to remove the tag from the second group and to add the tag to the third group upon the determining that the number of times that the tag has been assigned to the asset falls below a second threshold,

wherein the first threshold comprises a minimum number of times for the tag to be assigned to the asset in order for the tag to be considered one of the frequently-assigned tags,

wherein the second threshold comprises a maximum number of times for the tag to be assigned to the asset in order for the tag to be considered one of the rarely-assigned tags.

7. The computer-based system of claim 6, wherein the first group of tags comprises a first percentage of tags from the plurality of tags having a highest frequency of being assigned to the asset, wherein the third group of tags comprises a second percentage of tags from the plurality of tags having a lowest frequency of being assigned to the asset, and wherein the second a group of tags comprises a remaining percentage of tags between the first group of tags and the third group of tags.

8. The computer-based system of claim 6, wherein the tag suggestion module is further Currently Amended to randomly or pseudo-randomly select the sample of the long-tail tags.

9. The computer-based system of claim 6, wherein each of the tags comprises one or more keywords.

10. A non-transitory computer-readable medium having computer-executable instructions stored thereon which, when executed by a computer, cause the computer to:

identify a first group of tags from a plurality of tags, the first group of tags comprising frequently-assigned tags having a higher frequency of being assigned to an asset;

identify a second a group of tags from the plurality of tags, the second group of tags comprising long-tail tags having a lower frequency of being assigned to the asset than the frequently-assigned tags;

identify a third group of tags from the plurality of tags, the third group of tags comprising rarely-assigned tags having a lower frequency of being assigned to the asset than the long-tail tags;

suggest the frequently-assigned tags and a sample of the long-tail tags to a user, while disregarding the rarely-assigned tags, upon receiving a request over a network from the user to the tag the asset;

track, over a given number of chances, a number of times that a tag in the second group has been assigned to the asset when the tag is suggested;

determine whether the number of times that the tag has been suggested to the user surpasses a first threshold;

wherein the first threshold comprises a minimum number of times for the tag to be assigned to the asset in order for the tag to be considered one of the frequently-assigned tags,

upon determining that the number of times that the tag has been suggested to the user surpasses the first threshold, remove the tag from the second group of tags and add the tag into the first group of tags;

determine whether the number of times that the tag has been suggested to the user falls below a second threshold;

wherein the second threshold comprises a maximum number of times for the tag to be assigned to the asset in order for the tag to be considered one of the rarely-assigned tags, and

upon determining that the number of times that the tag has been suggested to the user falls below the second threshold, remove the tag from the second group of tags and add the tag into the third group of tags.

11. The computer-readable medium of claim 10, wherein to suggest the frequently-assigned tags and a sample of the long-tail tags to a user upon receiving a request over the network from the user to tag the asset, the computer-readable medium has further computer-executable instructions stored thereon which, when executed by a computer, cause the computer to:

select a given number of most frequently-assigned tags; and

to

randomly or pseudo-randomly select the sample of the long-tail tags.

12. The computer-readable medium of claim 10, wherein each of the tags comprises one or more keywords.