Taxonomy + Folksonomy ?

27 Apr 2007 - 9:12am
7 years ago
7 replies
1314 reads
Christopher Fahey
2005

Howdy all,

(sorry for cross-posting, I don't know where to go anymore)

I'm imagining a system (for a project) where a fixed taxonomy exists,
but where each node in the taxonomy is also associatated (editorially,
by humans) with a series of folksonomy tags.

So, for example, the system's fixed taxonomy might have:

Animals-->Carnivores-->Felines-->DomesticCats

Each node of this taxonomy might then be associated with various free
folksonomic tags, so the taxonomic "Animals" might also include any data
object tagged with the folksonomy tag "critters". Objects tagged "meat
eaters" or "predators" would fall under "Carnivores"

As users create arbitrary new folksonomy tags (for example someone
spontaneously tags an article about housecats with "kitties"), an editor
will catch the new tag and then find a close analogue among the fixed
taxonomy (in this case, "kitties would map to "DomesticCats").

This system implies that much of the content will not marked by the
taxonomy, either because it is being imported from elsewhere or because
we don't want to show a complex taxonomy UI to a user, instead just
letting them tag freely.

Has anyone seen a system like this in place somewhere?

Is this simply a synonym system, or is it something more?

Any thoughts on such a system?

Thanks,
-Cf

Christopher Fahey
____________________________
Behavior
http://www.behaviordesign.com
212.532.4002 x203
646.338.4002 mobile

Comments

27 Apr 2007 - 9:46am
Kyle Cooney
2006

Hi, Christopher -

I believe Buzzillions.com implements a system similar to what you're
describing to great effect:

http://www.buzzillions.com/ResultsController.do?N=4294967010

I've seen similar sites, but this is only one off the top of my head.

As for my thoughts on such a system, my opinion is that it's probably the
optimal solution for most cases. I've worked with people on both ends of
the spectrum, one being that the taxonomy should be completely fixed (no
user submissions), and one where the person thought the system should be
completely user defined, and that algorithms could be developed to take
advantage of the network effect for accuracy.

kyle

On 4/27/07, Christopher Fahey <chris.fahey at behaviordesign.com> wrote:
>
> Howdy all,
>
> (sorry for cross-posting, I don't know where to go anymore)
>
> I'm imagining a system (for a project) where a fixed taxonomy exists,
> but where each node in the taxonomy is also associatated (editorially,
> by humans) with a series of folksonomy tags.
>
> So, for example, the system's fixed taxonomy might have:
>
> Animals-->Carnivores-->Felines-->DomesticCats
>
> Each node of this taxonomy might then be associated with various free
> folksonomic tags, so the taxonomic "Animals" might also include any data
> object tagged with the folksonomy tag "critters". Objects tagged "meat
> eaters" or "predators" would fall under "Carnivores"
>
> As users create arbitrary new folksonomy tags (for example someone
> spontaneously tags an article about housecats with "kitties"), an editor
> will catch the new tag and then find a close analogue among the fixed
> taxonomy (in this case, "kitties would map to "DomesticCats").
>
> This system implies that much of the content will not marked by the
> taxonomy, either because it is being imported from elsewhere or because
> we don't want to show a complex taxonomy UI to a user, instead just
> letting them tag freely.
>
> Has anyone seen a system like this in place somewhere?
>
> Is this simply a synonym system, or is it something more?
>
> Any thoughts on such a system?
>
> Thanks,
> -Cf
>
> Christopher Fahey
> ____________________________
> Behavior
> http://www.behaviordesign.com
> 212.532.4002 x203
> 646.338.4002 mobile
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>

--
Of Interest: http://del.icio.us/kccdc
ph: 202.270.8214
2505 13th Street NW Apt #411
Washington, DC
20009

27 Apr 2007 - 9:49am
Thomas Vander Wal
2004

Chris --

This is getting to be a normal request (from my side of things).
Taxonomies are resource intensive to build and maintaing and
folksonomies lack structure. But, when you start tying the two
together you start to get a taxonomy that is emergent and is
inclusive. Many who are seeking information can not find it because
the taxonomy is not in their terminology/vocabulary (this happens a
lot).

What you are doing in your example are using folksonomies to validate
your taxonomy, identify gaps, and provide the terms to fill those
gaps. The gaps may be sectors or disciplines that were missed
(errantly, or more than likely did not register as a large enough
population to address) or it may be emergent terminology/vocabulary.

This is a practice that has been providing a decent chunk of my
advising the past year or so. There are not tools that do this as of
yet, but that should be addressed in the near future.

All the best,
Thomas

On 4/27/07, Christopher Fahey <chris.fahey at behaviordesign.com> wrote:
> Howdy all,
>
> (sorry for cross-posting, I don't know where to go anymore)
>
> I'm imagining a system (for a project) where a fixed taxonomy exists,
> but where each node in the taxonomy is also associatated (editorially,
> by humans) with a series of folksonomy tags.
>
> So, for example, the system's fixed taxonomy might have:
>
> Animals-->Carnivores-->Felines-->DomesticCats
>
> Each node of this taxonomy might then be associated with various free
> folksonomic tags, so the taxonomic "Animals" might also include any data
> object tagged with the folksonomy tag "critters". Objects tagged "meat
> eaters" or "predators" would fall under "Carnivores"
>
> As users create arbitrary new folksonomy tags (for example someone
> spontaneously tags an article about housecats with "kitties"), an editor
> will catch the new tag and then find a close analogue among the fixed
> taxonomy (in this case, "kitties would map to "DomesticCats").
>
> This system implies that much of the content will not marked by the
> taxonomy, either because it is being imported from elsewhere or because
> we don't want to show a complex taxonomy UI to a user, instead just
> letting them tag freely.
>
> Has anyone seen a system like this in place somewhere?
>
> Is this simply a synonym system, or is it something more?
>
> Any thoughts on such a system?
>
> Thanks,
> -Cf
>
> Christopher Fahey
> ____________________________
> Behavior
> http://www.behaviordesign.com
> 212.532.4002 x203
> 646.338.4002 mobile
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>

27 Apr 2007 - 10:49am
lachica
2006

This seems like something that will become more automated in the future.
Take similicio.us (http://similicio.us/info.html), for example. It searches
sites tagged in del.icio.us and uses them to generate lists of similar sites
based on associated sites. Are there any association engines out there that
are generating more formal taxonomies from informal folksonomies?

Best,
Julie

On 4/27/07, Thomas Vander Wal <vanderwal at gmail.com> wrote:
>
> Chris --
>
> This is getting to be a normal request (from my side of things).
> Taxonomies are resource intensive to build and maintaing and
> folksonomies lack structure. But, when you start tying the two
> together you start to get a taxonomy that is emergent and is
> inclusive. Many who are seeking information can not find it because
> the taxonomy is not in their terminology/vocabulary (this happens a
> lot).
>
> What you are doing in your example are using folksonomies to validate
> your taxonomy, identify gaps, and provide the terms to fill those
> gaps. The gaps may be sectors or disciplines that were missed
> (errantly, or more than likely did not register as a large enough
> population to address) or it may be emergent terminology/vocabulary.
>
> This is a practice that has been providing a decent chunk of my
> advising the past year or so. There are not tools that do this as of
> yet, but that should be addressed in the near future.
>
> All the best,
> Thomas
>
> On 4/27/07, Christopher Fahey <chris.fahey at behaviordesign.com> wrote:
> > Howdy all,
> >
> > (sorry for cross-posting, I don't know where to go anymore)
> >
> > I'm imagining a system (for a project) where a fixed taxonomy exists,
> > but where each node in the taxonomy is also associatated (editorially,
> > by humans) with a series of folksonomy tags.
> >
> > So, for example, the system's fixed taxonomy might have:
> >
> > Animals-->Carnivores-->Felines-->DomesticCats
> >
> > Each node of this taxonomy might then be associated with various free
> > folksonomic tags, so the taxonomic "Animals" might also include any data
> > object tagged with the folksonomy tag "critters". Objects tagged "meat
> > eaters" or "predators" would fall under "Carnivores"
> >
> > As users create arbitrary new folksonomy tags (for example someone
> > spontaneously tags an article about housecats with "kitties"), an editor
> > will catch the new tag and then find a close analogue among the fixed
> > taxonomy (in this case, "kitties would map to "DomesticCats").
> >
> > This system implies that much of the content will not marked by the
> > taxonomy, either because it is being imported from elsewhere or because
> > we don't want to show a complex taxonomy UI to a user, instead just
> > letting them tag freely.
> >
> > Has anyone seen a system like this in place somewhere?
> >
> > Is this simply a synonym system, or is it something more?
> >
> > Any thoughts on such a system?
> >
> > Thanks,
> > -Cf
> >
> > Christopher Fahey
> > ____________________________
> > Behavior
> > http://www.behaviordesign.com
> > 212.532.4002 x203
> > 646.338.4002 mobile
> > ________________________________________________________________
> > Welcome to the Interaction Design Association (IxDA)!
> > To post to this list ....... discuss at ixda.org
> > List Guidelines ............ http://listguide.ixda.org/
> > List Help .................. http://listhelp.ixda.org/
> > (Un)Subscription Options ... http://subscription-options.ixda.org/
> > Announcements List ......... http://subscribe-announce.ixda.org/
> > Questions .................. lists at ixda.org
> > Home ....................... http://ixda.org/
> > Resource Library ........... http://resources.ixda.org
> >
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>

29 May 2007 - 8:43am
itst
2007

Julie,

there are two approaches to generate clusters and hierarchies from
folksonomies I know of.

[Begelman2006]
Begelman, Grigory; Keller, Philipp & Smadja, Frank: Automated Tag
Clustering: Improving search and exploration in the tag space. In the
Proceedings of WWW 2006 Collaborative Web. Tagging Workshop,
Edinburgh, Scotland. 2006.
http://www.rawsugar.com/www2006/automatedTagClustering.pdf

[Heymann2006]
Heymann, Paul & Garcia-Molina, Hector: Collaborative Creation of
Communal Hierarchical Taxonomies in Social Tagging Systems.
Department of Computer Science, Stanford University, Stanford, CA,
USA, April 2006.
http://dbpubs.stanford.edu:8090/pub/2006-10

Begelman, Keller and Smadja try to create tag clusters based on the
frequency of occurrence of tags in the same tag group (a tag group is
the entirety of tags used by one user to describe one resource).

Heymann and Garcia-Molina describe a system that generates a
hierarchy based on more or less the same similarity Begelman, Keller
and Smadja use. But instead of creating clusters, they compute the
similarity and use this value to either order a tag under another one
take this tag and treat it as a new heading.

hth,
Sascha

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://beta.ixda.org/discuss?post=16143

29 May 2007 - 9:24am
jayeffvee
2007

I'm getting a "file not found" message from the first link...

-----Original Message-----
From: discuss-bounces at lists.interactiondesigners.com
[mailto:discuss-bounces at lists.interactiondesigners.com] On Behalf Of
Sascha A.Carlin
Sent: Tuesday, May 29, 2007 9:44 AM
To: discuss at lists.interactiondesigners.com
Subject: Re: [IxDA Discuss] Taxonomy Folksonomy ?

Julie,

there are two approaches to generate clusters and hierarchies from
folksonomies I know of.

[Begelman2006]
Begelman, Grigory; Keller, Philipp & Smadja, Frank: Automated Tag
Clustering: Improving search and exploration in the tag space. In the
Proceedings of WWW 2006 Collaborative Web. Tagging Workshop,
Edinburgh, Scotland. 2006.
http://www.rawsugar.com/www2006/automatedTagClustering.pdf

[Heymann2006]
Heymann, Paul & Garcia-Molina, Hector: Collaborative Creation of
Communal Hierarchical Taxonomies in Social Tagging Systems.
Department of Computer Science, Stanford University, Stanford, CA,
USA, April 2006.
http://dbpubs.stanford.edu:8090/pub/2006-10

Begelman, Keller and Smadja try to create tag clusters based on the
frequency of occurrence of tags in the same tag group (a tag group is
the entirety of tags used by one user to describe one resource).

Heymann and Garcia-Molina describe a system that generates a
hierarchy based on more or less the same similarity Begelman, Keller
and Smadja use. But instead of creating clusters, they compute the
similarity and use this value to either order a tag under another one
take this tag and treat it as a new heading.

hth,
Sascha

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://beta.ixda.org/discuss?post=16143

________________________________________________________________
Welcome to the Interaction Design Association (IxDA)!
To post to this list ....... discuss at ixda.org
List Guidelines ............ http://listguide.ixda.org/
List Help .................. http://listhelp.ixda.org/
(Un)Subscription Options ... http://subscription-options.ixda.org/
Announcements List ......... http://subscribe-announce.ixda.org/
Questions .................. lists at ixda.org
Home ....................... http://ixda.org/
Resource Library ........... http://resources.ixda.org

29 May 2007 - 9:38am
Will Parker
2007

Trust Google (but verify).

A Google search on "Automated Tag Clustering" turns up the file as
the first hit, probably because this is just the sort of thing Larry
and Sergey would want to see. };->

Here's the correct link at RawSugar (also courtesy Google):

http://www.rawsugar.com/www2006/20.pdf

- Will

Will Parker
wparker at ChannelingDesign.com

“I wish developing great products was as easy as writing a check. If
that were the case, then Microsoft would have great products.” -
Steve Jobs

On May 29, 2007, at 7:24 AM, Vermette, Joan wrote:

> I'm getting a "file not found" message from the first link...
>
> -----Original Message-----
> From: discuss-bounces at lists.interactiondesigners.com
> [mailto:discuss-bounces at lists.interactiondesigners.com] On Behalf Of
> Sascha A.Carlin
> Sent: Tuesday, May 29, 2007 9:44 AM
> To: discuss at lists.interactiondesigners.com
> Subject: Re: [IxDA Discuss] Taxonomy Folksonomy ?
>
> Julie,
>
> there are two approaches to generate clusters and hierarchies from
> folksonomies I know of.
>
> [Begelman2006]
> Begelman, Grigory; Keller, Philipp & Smadja, Frank: Automated Tag
> Clustering: Improving search and exploration in the tag space. In the
> Proceedings of WWW 2006 Collaborative Web. Tagging Workshop,
> Edinburgh, Scotland. 2006.
> http://www.rawsugar.com/www2006/automatedTagClustering.pdf
>
> [Heymann2006]
> Heymann, Paul & Garcia-Molina, Hector: Collaborative Creation of
> Communal Hierarchical Taxonomies in Social Tagging Systems.
> Department of Computer Science, Stanford University, Stanford, CA,
> USA, April 2006.
> http://dbpubs.stanford.edu:8090/pub/2006-10
>
> Begelman, Keller and Smadja try to create tag clusters based on the
> frequency of occurrence of tags in the same tag group (a tag group is
> the entirety of tags used by one user to describe one resource).
>
> Heymann and Garcia-Molina describe a system that generates a
> hierarchy based on more or less the same similarity Begelman, Keller
> and Smadja use. But instead of creating clusters, they compute the
> similarity and use this value to either order a tag under another one
> take this tag and treat it as a new heading.
>
> hth,
> Sascha
>
>
> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
> Posted from the new ixda.org
> http://beta.ixda.org/discuss?post=16143
>
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org

29 May 2007 - 10:00am
itst
2007

Joan,

> I'm getting a "file not found" message from the first link...

sorry, I copy&pasted the link without checking it. It seems like the document
is no longer available, since RawSugar closed its doors last year. And it's not
included in the proceedings of WWW2006...

The best I can come up with right now is this document:
http://www.pui.ch/phred/automated_tag_clustering/

--
Sascha

Syndicate content Get the feed