SlideShare a Scribd company logo
1 of 30
Tag Gardening Activities for Folksonomy
     Maintenance and Enrichment



     Katrin Weller & Isabella Peters
       Heinrich-Heine-University Düsseldorf
      Institute for Language and Information
            Dept. of Information Science

 Presentation at I-Semantics, Triple-I Conference 2008.
              Graz, 04. September 2008
Motivation


    How to combine the dynamics of freely chosen
     How to combine the dynamics of freely chosen
    tags with the steadiness and complexity of
     tags with the steadiness and complexity of
    controlled vocabularies?
     controlled vocabularies?
Two Sides of the Same Coin


• Aim 1: Optimizing folksonomies for specific application
  scenarios.

• Aim 2: Maintaining and enriching knowledge organization
  systems (KOS, e.g. ontologies or thesauri) with
  folksonomy terms.

  Tag Gardening!
The Two Sides of Tag Gardening
Some Problems of Unstructured
Folksonomies

• Spelling variants, translations, abbreviations and
  synonyms have to be considered for query formulation
  and indexing.

• Tags serve a variety of functions, not just content
  description (e.g. “toread”, “@Henry”).

• Tags enable browsing by popularity – but not navigation
  by meanings and semantic interrelations.

• Even on a personal level unstructured tags may become
  unmanageable.
Source:
http://flickr.com/photos/jup3nep/2473905053/
Some problems of unstructured
folksonomies




    Tag cloud of the user„MarmaladeToday’s“ for
    Tag cloud of the user„MarmaladeToday’s“ for
    1.157 bookmarks. Source: del.icio.us (5.03.2008).
    1.157 bookmarks. Source: del.icio.us (5.03.2008).
Some approaches:
Adding structure to folksonomies




            Del.icio.us
            Del.icio.us




                                   Bibsonomy.org
                                   Bibsonomy.org
Some solutions:
Adding structure to folksonomies




                                   Flickr.com
                                    Flickr.com
The Folksonomy Tag Garden
Tag Gardening

• Term coined by James Governor:
  „Like plants or animals, tags evolve in an emergent
  fashion, open to hybridisation. Stewardship can help grow
  and put roots down. Helping the darwinian process is tag
  gardening.“

• We define tag gardening as any activity to edit,
  reeingineer, manipulate or organize tags – in order to
  make them more productive and effective.

  Tag gardening is performed on top of existing
  folksonomies. Users may tag as usual, afterwards
  gardenig activities may be performed for optimization
  and user support.
Levels of Tag Gardening
Document collection vs. single document level:
  Simple form: editing the tags of one single document.
  Complex task: handling all tags in a folksonomy system.

Personal vs. collaborative level:
  Personomy level: single users edit the personal tags they
  use within a system.
  Folksonomy level: enabling a user community to
  collaboratively maintain all tags in use.

Intra- and cross-platform level:
   Usual case: consider only tags within one platform.
   Broader view: for some cases the use of consistent tags
   across different platforms will be useful.
Tag Gardening Activities:
 1. Weeding

• Removing “bad tags“: spelling variants (plural vs. singular,
  conflation of multi-word tags) and spam through
  “pesticides“.
• Aim: enhancing recall and a consistent indexing
  vocabulary.
• Achieved by
   – type-ahead functionality during indexing,
   – editing functionalities for tags after the application (remove,
     change, etc.),
   – Natural Language Processing of index tags and search tags,
   – indexing and retrieval tutorials or guidelines for users,
   – authorized users as gardeners
  Simplest form of tag gardening
Tag Gardening Activities:
2. Seeding
• Extending the folksonomy with rarely used “seedlings“
  if high-frequency tags do not sufficiently discriminate
  resources.
• Aim: enhancing precision and expressiveness of the
  folksonomy.
• Achieved by
   – displaying an inverse tag cloud during indexing or
     particular “green house“ areas where the seedlings may
     develop and grow,
   – discrete tag suggestions during indexing.
Tag Green Houses


• Problem: high-frequent tags provide high recall – but low
  precision and low degrees of discrimination.



                                               Tag „Design“ on
                                                Tag „Design“ on
                                               Del.icio.us.
                                                Del.icio.us.
• Seedlings or „Baby tags“ as additional entry points to
  explore document collections.
• Promoting baby tags via alternative display options like
   – „New tags“ / trends
   – „infrequent tags“ / inverse tag cloud
Tag Gardening Activities:
  3. Landscape Architecture

• Shaping the folksonomy into “flower beds“, distinguishing
  similar looking “plants“, identifying their “species“,
  branding each species with labels and giving additional
  information regarding their application area.
• Aim: enhancing precision and expressiveness of the
  folksonomy by adding semantics. For query expansion
  along semantic relations, for enhanced navigation, as basis
  for semantic-oriented display.
• Achieved by
   –   conflation of multi-language tags,
   –   summarization of synonyms,
   –   distinction of homonyms,
   –   establishment of semantic relations,
   –   field-based tagging
Synonym interrelation and distinction of homonyms.
Hidden Semantic Relations
      in Folksonomies




Source for image: http:// www.flickr.com
Tag Gardening Activities:
4. Fertilizing

• Combination of folksonomies and KOS during indexing
  and retrieval.
• Aim: enhancing precision and recall and the
  expressiveness of the folksonomies by adding
  semantics, for query expansion during retrieval via
  semantic relations, for enhanced indexing
  functionalities, for enhanced navigation within the
  folksonomy, as basis for semantic-oriented displays.
• Achieved by
   – semantic-oriented tag suggestions during indexing and
     retrieval ( tag suggestions not based on tag popularity
     to avoid “success breeds success-effect”)
   – establishment of semantic relations by mappings to KOS
     (after indexing)
Tag Gardening Activities:
Fertilizing Type 1 and 2



• Fertilizing Type 1
  Fertilizing a folksonomy with semantic structures from
  other knowledge organization systems.



• Fertilizing Type 2
  Fertilizing a structured knowledge organization system
  (ontology, thesaurus) with user-generated terminologies.
Tag Gardening in Practice


Possibilities
   – Editing and deleting functionalities for tags on
     personal and document level.
   – Detecting and labeling semantic relations.
   – Use of power tags as candidates on document
     collection level.
   – Co-occurence computations as starting points.
   – Tools for personal tag gardening on personomy level.
   – Collaborative tag gardening in (small) communities.
   – Power users as chief gardeners in communities.
Gardening support: Power tags and co-
occurences as candidates and starting points

Steps:

• Detecting power tags on the resource
  level
• Computing co-occurences of power tags
  with the whole tag collection
• Result: distribution and power tags on co-
  occurence level


  Power relations as candidates for new
  folksonomy structuring approaches

                                               Data retrieved 2008-05-15
                                               from del.icio.us
Personal Tag Gardening

• Our approach: TagCare
  (www.tagcare.org, currently under development)

• „take care of your tags“
  TagCare imports personal tags from different tagging
  platforms and allows to manage them centrally for better
  personal overview.

Personal structured tag vocabulary for:
• Personal infomation management (PIM)
• Consistent tagging across platforms (functionalities under
  development).
• Search in folksonomy-based systems (e.g. use pre-
  cooked personal synonym lists).
Personal Tag Gardening


Use of TagCare – first version


•    Adding, deleting and editing of tags
•    Hierarchical structuring of tags
•    Interrelating of synonyms
•    Labeling of otherwise related tags
• Statistics on tag frequencies
Tag Gardening Community

Solution for small, closed communties and
development of shared vocabularies:
Community tagging plus single information
architects.
                                   Information architect:
                                    Information architect:
                                      builds a structured thesaurus
                                       builds a structured thesaurus
                                   and enhances it with tags.
                                    and enhances it with tags.
                                      establishes structures between
                                       establishes structures between
                                   loose tags (emergent semantics).
                                    loose tags (emergent semantics).



                                        Community:
                                         Community:
                                           applies tags and
                                            applies tags and
                                        performs minor editing
                                         performs minor editing
                                        activities.
                                         activities.
Collaborative Tag Gardening


Challenges and open questions
   – Who may perform editing actions? Who is an
     authorized user?
   – Which tags are spam?
   – How can users collaborate, which aspects are
     determined collectively?
   – Can tagging guidelines be applied?
   – …
Challenges and open questions

• The three elements of folksonomies (resources – people –
  tags) constitute a three-dimensional knowledge space:
  domain – community – expressiveness.




• Conflicts are caused, if all three elements should be
  considered to large extents simultaneously.
Conclusions & Outlook

• Folksonomies may be enriched with semantics to achieve
  a combination of vocabulary dynamics and structure.
• Aim: improved information retrieval on personomy and
  collection level.
• Manual and semi-automatic approaches may be
  combined.
• Tag gardening may first be applied to small communities
  or small application domains.
• Solutions for bigger communities are needed, „chief
  gardeners“ as first approach.

• Future work: development of TagCare, analysis of types
  of semantic relations for folksonomy enrichment, tagging
  for support of ontology or thesaurus development.
Thank You

  Best regards from Düsseldorf!




Contact:   katrin.weller@uni-duesseldorf.de
           isabella.peters@uni-duesseldorf.de
           http://www.phil-fak.uni-duesseldorf.de/infowiss/

More Related Content

Similar to Seeding Weeding Fertilizing - Tag Gardening for Folksonomy Maintenance

Indexing presentation 2013 06-04
Indexing presentation 2013 06-04Indexing presentation 2013 06-04
Indexing presentation 2013 06-04
Louise Spiteri
 
User-generated metadata: Boon or bust for indexing and controlled vocabularies?
User-generated metadata: Boon or bust for indexing and controlled vocabularies?User-generated metadata: Boon or bust for indexing and controlled vocabularies?
User-generated metadata: Boon or bust for indexing and controlled vocabularies?
Louise Spiteri
 
User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?
User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?
User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?
Louise Spiteri
 
Tag And Tag Based Recommender
Tag And Tag Based RecommenderTag And Tag Based Recommender
Tag And Tag Based Recommender
gu wendong
 
Tagging For Community of Practice
Tagging For Community of PracticeTagging For Community of Practice
Tagging For Community of Practice
Peter Rawsthorne
 
Social taggingpresentation
Social taggingpresentationSocial taggingpresentation
Social taggingpresentation
molleem
 
int.ere.st: SCOT-based Tag Sharing Services
int.ere.st: SCOT-based Tag Sharing Servicesint.ere.st: SCOT-based Tag Sharing Services
int.ere.st: SCOT-based Tag Sharing Services
Haklae Kim
 

Similar to Seeding Weeding Fertilizing - Tag Gardening for Folksonomy Maintenance (20)

IMT530 Tagging Presentation
IMT530 Tagging PresentationIMT530 Tagging Presentation
IMT530 Tagging Presentation
 
Folksonomy
FolksonomyFolksonomy
Folksonomy
 
Indexing presentation 2013 06-04
Indexing presentation 2013 06-04Indexing presentation 2013 06-04
Indexing presentation 2013 06-04
 
User-generated metadata: Boon or bust for indexing and controlled vocabularies?
User-generated metadata: Boon or bust for indexing and controlled vocabularies?User-generated metadata: Boon or bust for indexing and controlled vocabularies?
User-generated metadata: Boon or bust for indexing and controlled vocabularies?
 
User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?
User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?
User-Generated Metadata: Boon or Bust for Indexing and Controlled Vocabularies?
 
Tag And Tag Based Recommender
Tag And Tag Based RecommenderTag And Tag Based Recommender
Tag And Tag Based Recommender
 
FaceTag at IASummit 2007
FaceTag at IASummit 2007FaceTag at IASummit 2007
FaceTag at IASummit 2007
 
FaceTag - IASummit 2007
FaceTag - IASummit 2007FaceTag - IASummit 2007
FaceTag - IASummit 2007
 
Folksonomies & social tagging
Folksonomies & social taggingFolksonomies & social tagging
Folksonomies & social tagging
 
Folksonomy and Tagging in the Social Web
Folksonomy and Tagging in the Social WebFolksonomy and Tagging in the Social Web
Folksonomy and Tagging in the Social Web
 
Using Controlled Vocabularies
Using Controlled VocabulariesUsing Controlled Vocabularies
Using Controlled Vocabularies
 
FaceTag: Integrating Bottom-up and Top-down Classification in a Social Taggin...
FaceTag: Integrating Bottom-up and Top-down Classification in a Social Taggin...FaceTag: Integrating Bottom-up and Top-down Classification in a Social Taggin...
FaceTag: Integrating Bottom-up and Top-down Classification in a Social Taggin...
 
Hybrid Approaches to Taxonomy & Folksonmy
Hybrid Approaches to Taxonomy & FolksonmyHybrid Approaches to Taxonomy & Folksonmy
Hybrid Approaches to Taxonomy & Folksonmy
 
Dynamic Potential of Semantic Enrichment
Dynamic Potential of Semantic EnrichmentDynamic Potential of Semantic Enrichment
Dynamic Potential of Semantic Enrichment
 
Exploring Social Bookmarking
Exploring Social BookmarkingExploring Social Bookmarking
Exploring Social Bookmarking
 
Some thoughts on social tagging
Some thoughts on social taggingSome thoughts on social tagging
Some thoughts on social tagging
 
Semantics To The Bookmarks: A Review of Social Semantic Bookmarking Systems
Semantics To The Bookmarks: A Review of Social Semantic Bookmarking SystemsSemantics To The Bookmarks: A Review of Social Semantic Bookmarking Systems
Semantics To The Bookmarks: A Review of Social Semantic Bookmarking Systems
 
Tagging For Community of Practice
Tagging For Community of PracticeTagging For Community of Practice
Tagging For Community of Practice
 
Social taggingpresentation
Social taggingpresentationSocial taggingpresentation
Social taggingpresentation
 
int.ere.st: SCOT-based Tag Sharing Services
int.ere.st: SCOT-based Tag Sharing Servicesint.ere.st: SCOT-based Tag Sharing Services
int.ere.st: SCOT-based Tag Sharing Services
 

More from Katrin Weller

Quantität vor Qualität? Big Data im Kontext von Social Media Daten
Quantität vor Qualität? Big Data im Kontext von Social Media DatenQuantität vor Qualität? Big Data im Kontext von Social Media Daten
Quantität vor Qualität? Big Data im Kontext von Social Media Daten
Katrin Weller
 

More from Katrin Weller (20)

Weller pleasures+perils social media
Weller pleasures+perils social mediaWeller pleasures+perils social media
Weller pleasures+perils social media
 
Weller social media as research data_psm15
Weller social media as research data_psm15Weller social media as research data_psm15
Weller social media as research data_psm15
 
Fail ir16 intro
Fail ir16 introFail ir16 intro
Fail ir16 intro
 
Fail! workshop introduction at Web Science Conference
Fail! workshop introduction at Web Science ConferenceFail! workshop introduction at Web Science Conference
Fail! workshop introduction at Web Science Conference
 
Challenges in-archiving-twitter
Challenges in-archiving-twitterChallenges in-archiving-twitter
Challenges in-archiving-twitter
 
The digital traces of user generated content
The digital traces of user generated contentThe digital traces of user generated content
The digital traces of user generated content
 
The Hidden Data of Social Media Rearch_CSS-winter-symposium
The Hidden Data of Social Media Rearch_CSS-winter-symposiumThe Hidden Data of Social Media Rearch_CSS-winter-symposium
The Hidden Data of Social Media Rearch_CSS-winter-symposium
 
Twitter-Daten in der sozialwissenschaftlichen Forschung – Möglichkeiten und H...
Twitter-Daten in der sozialwissenschaftlichen Forschung – Möglichkeiten und H...Twitter-Daten in der sozialwissenschaftlichen Forschung – Möglichkeiten und H...
Twitter-Daten in der sozialwissenschaftlichen Forschung – Möglichkeiten und H...
 
Publishing with impact
Publishing with impactPublishing with impact
Publishing with impact
 
"I always feel it must be great to be a hacker"
"I always feel it must be great to be a hacker" "I always feel it must be great to be a hacker"
"I always feel it must be great to be a hacker"
 
Social-Media-Forschung
Social-Media-ForschungSocial-Media-Forschung
Social-Media-Forschung
 
Hidden Data of Social Media Research
Hidden Data of Social Media ResearchHidden Data of Social Media Research
Hidden Data of Social Media Research
 
Big data - Gewinnung, Auswertung und Darstellung großer Mengen onlinegenerier...
Big data - Gewinnung, Auswertung und Darstellung großer Mengen onlinegenerier...Big data - Gewinnung, Auswertung und Darstellung großer Mengen onlinegenerier...
Big data - Gewinnung, Auswertung und Darstellung großer Mengen onlinegenerier...
 
What’s new in social media research?
What’s new in social media research?What’s new in social media research?
What’s new in social media research?
 
Twitter-Daten in der sozialwissenschaftlichen Forschung
Twitter-Daten in der sozialwissenschaftlichen ForschungTwitter-Daten in der sozialwissenschaftlichen Forschung
Twitter-Daten in der sozialwissenschaftlichen Forschung
 
Social Media Research Methods
Social Media Research MethodsSocial Media Research Methods
Social Media Research Methods
 
Quantität vor Qualität? Big Data im Kontext von Social Media Daten
Quantität vor Qualität? Big Data im Kontext von Social Media DatenQuantität vor Qualität? Big Data im Kontext von Social Media Daten
Quantität vor Qualität? Big Data im Kontext von Social Media Daten
 
The pleasures and perils of studying Twitter
The pleasures and perils of studying TwitterThe pleasures and perils of studying Twitter
The pleasures and perils of studying Twitter
 
Friends or Followers. German Soccer Clubs and Their Fans on Twitter
Friends or Followers. German Soccer Clubs and Their Fans on TwitterFriends or Followers. German Soccer Clubs and Their Fans on Twitter
Friends or Followers. German Soccer Clubs and Their Fans on Twitter
 
What do we get from Twitter - and what not?
What do we get from Twitter - and what not?What do we get from Twitter - and what not?
What do we get from Twitter - and what not?
 

Recently uploaded

KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...
KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...
KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...
Cara Menggugurkan Kandungan 087776558899
 
Girls in Mahipalpur (delhi) call me [🔝9953056974🔝] escort service 24X7
Girls in Mahipalpur  (delhi) call me [🔝9953056974🔝] escort service 24X7Girls in Mahipalpur  (delhi) call me [🔝9953056974🔝] escort service 24X7
Girls in Mahipalpur (delhi) call me [🔝9953056974🔝] escort service 24X7
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 
February 2024 Recommendations for newsletter
February 2024 Recommendations for newsletterFebruary 2024 Recommendations for newsletter
February 2024 Recommendations for newsletter
ssuserdfec6a
 
the Husband rolesBrown Aesthetic Cute Group Project Presentation
the Husband rolesBrown Aesthetic Cute Group Project Presentationthe Husband rolesBrown Aesthetic Cute Group Project Presentation
the Husband rolesBrown Aesthetic Cute Group Project Presentation
brynpueblos04
 

Recently uploaded (15)

Emotional Freedom Technique Tapping Points Diagram.pdf
Emotional Freedom Technique Tapping Points Diagram.pdfEmotional Freedom Technique Tapping Points Diagram.pdf
Emotional Freedom Technique Tapping Points Diagram.pdf
 
2023 - Between Philosophy and Practice: Introducing Yoga
2023 - Between Philosophy and Practice: Introducing Yoga2023 - Between Philosophy and Practice: Introducing Yoga
2023 - Between Philosophy and Practice: Introducing Yoga
 
Social Learning Theory presentation.pptx
Social Learning Theory presentation.pptxSocial Learning Theory presentation.pptx
Social Learning Theory presentation.pptx
 
KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...
KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...
KLINIK BATA Jual obat penggugur kandungan 087776558899 ABORSI JANIN KEHAMILAN...
 
Exploring Stoic Philosophy From Ancient Wisdom to Modern Relevance.pdf
Exploring Stoic Philosophy From Ancient Wisdom to Modern Relevance.pdfExploring Stoic Philosophy From Ancient Wisdom to Modern Relevance.pdf
Exploring Stoic Philosophy From Ancient Wisdom to Modern Relevance.pdf
 
March 2023 Recommendations for newsletter
March 2023 Recommendations for newsletterMarch 2023 Recommendations for newsletter
March 2023 Recommendations for newsletter
 
Pokemon Go... Unraveling the Conspiracy Theory
Pokemon Go... Unraveling the Conspiracy TheoryPokemon Go... Unraveling the Conspiracy Theory
Pokemon Go... Unraveling the Conspiracy Theory
 
Goregaon West Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
Goregaon West Escorts 🥰 8617370543 Call Girls Offer VIP Hot GirlsGoregaon West Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
Goregaon West Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
 
SIKP311 Sikolohiyang Pilipino - Ginhawa.pptx
SIKP311 Sikolohiyang Pilipino - Ginhawa.pptxSIKP311 Sikolohiyang Pilipino - Ginhawa.pptx
SIKP311 Sikolohiyang Pilipino - Ginhawa.pptx
 
Dadar West Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
Dadar West Escorts 🥰 8617370543 Call Girls Offer VIP Hot GirlsDadar West Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
Dadar West Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
 
Girls in Mahipalpur (delhi) call me [🔝9953056974🔝] escort service 24X7
Girls in Mahipalpur  (delhi) call me [🔝9953056974🔝] escort service 24X7Girls in Mahipalpur  (delhi) call me [🔝9953056974🔝] escort service 24X7
Girls in Mahipalpur (delhi) call me [🔝9953056974🔝] escort service 24X7
 
February 2024 Recommendations for newsletter
February 2024 Recommendations for newsletterFebruary 2024 Recommendations for newsletter
February 2024 Recommendations for newsletter
 
Call Girls In Mumbai Just Genuine Call ☎ 7738596112✅ Call Girl Andheri East G...
Call Girls In Mumbai Just Genuine Call ☎ 7738596112✅ Call Girl Andheri East G...Call Girls In Mumbai Just Genuine Call ☎ 7738596112✅ Call Girl Andheri East G...
Call Girls In Mumbai Just Genuine Call ☎ 7738596112✅ Call Girl Andheri East G...
 
the Husband rolesBrown Aesthetic Cute Group Project Presentation
the Husband rolesBrown Aesthetic Cute Group Project Presentationthe Husband rolesBrown Aesthetic Cute Group Project Presentation
the Husband rolesBrown Aesthetic Cute Group Project Presentation
 
Colaba Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
Colaba Escorts 🥰 8617370543 Call Girls Offer VIP Hot GirlsColaba Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
Colaba Escorts 🥰 8617370543 Call Girls Offer VIP Hot Girls
 

Seeding Weeding Fertilizing - Tag Gardening for Folksonomy Maintenance

  • 1. Tag Gardening Activities for Folksonomy Maintenance and Enrichment Katrin Weller & Isabella Peters Heinrich-Heine-University Düsseldorf Institute for Language and Information Dept. of Information Science Presentation at I-Semantics, Triple-I Conference 2008. Graz, 04. September 2008
  • 2. Motivation How to combine the dynamics of freely chosen How to combine the dynamics of freely chosen tags with the steadiness and complexity of tags with the steadiness and complexity of controlled vocabularies? controlled vocabularies?
  • 3. Two Sides of the Same Coin • Aim 1: Optimizing folksonomies for specific application scenarios. • Aim 2: Maintaining and enriching knowledge organization systems (KOS, e.g. ontologies or thesauri) with folksonomy terms. Tag Gardening!
  • 4. The Two Sides of Tag Gardening
  • 5. Some Problems of Unstructured Folksonomies • Spelling variants, translations, abbreviations and synonyms have to be considered for query formulation and indexing. • Tags serve a variety of functions, not just content description (e.g. “toread”, “@Henry”). • Tags enable browsing by popularity – but not navigation by meanings and semantic interrelations. • Even on a personal level unstructured tags may become unmanageable.
  • 7. Some problems of unstructured folksonomies Tag cloud of the user„MarmaladeToday’s“ for Tag cloud of the user„MarmaladeToday’s“ for 1.157 bookmarks. Source: del.icio.us (5.03.2008). 1.157 bookmarks. Source: del.icio.us (5.03.2008).
  • 8. Some approaches: Adding structure to folksonomies Del.icio.us Del.icio.us Bibsonomy.org Bibsonomy.org
  • 9. Some solutions: Adding structure to folksonomies Flickr.com Flickr.com
  • 11. Tag Gardening • Term coined by James Governor: „Like plants or animals, tags evolve in an emergent fashion, open to hybridisation. Stewardship can help grow and put roots down. Helping the darwinian process is tag gardening.“ • We define tag gardening as any activity to edit, reeingineer, manipulate or organize tags – in order to make them more productive and effective. Tag gardening is performed on top of existing folksonomies. Users may tag as usual, afterwards gardenig activities may be performed for optimization and user support.
  • 12. Levels of Tag Gardening Document collection vs. single document level: Simple form: editing the tags of one single document. Complex task: handling all tags in a folksonomy system. Personal vs. collaborative level: Personomy level: single users edit the personal tags they use within a system. Folksonomy level: enabling a user community to collaboratively maintain all tags in use. Intra- and cross-platform level: Usual case: consider only tags within one platform. Broader view: for some cases the use of consistent tags across different platforms will be useful.
  • 13. Tag Gardening Activities: 1. Weeding • Removing “bad tags“: spelling variants (plural vs. singular, conflation of multi-word tags) and spam through “pesticides“. • Aim: enhancing recall and a consistent indexing vocabulary. • Achieved by – type-ahead functionality during indexing, – editing functionalities for tags after the application (remove, change, etc.), – Natural Language Processing of index tags and search tags, – indexing and retrieval tutorials or guidelines for users, – authorized users as gardeners Simplest form of tag gardening
  • 14. Tag Gardening Activities: 2. Seeding • Extending the folksonomy with rarely used “seedlings“ if high-frequency tags do not sufficiently discriminate resources. • Aim: enhancing precision and expressiveness of the folksonomy. • Achieved by – displaying an inverse tag cloud during indexing or particular “green house“ areas where the seedlings may develop and grow, – discrete tag suggestions during indexing.
  • 15. Tag Green Houses • Problem: high-frequent tags provide high recall – but low precision and low degrees of discrimination. Tag „Design“ on Tag „Design“ on Del.icio.us. Del.icio.us. • Seedlings or „Baby tags“ as additional entry points to explore document collections. • Promoting baby tags via alternative display options like – „New tags“ / trends – „infrequent tags“ / inverse tag cloud
  • 16. Tag Gardening Activities: 3. Landscape Architecture • Shaping the folksonomy into “flower beds“, distinguishing similar looking “plants“, identifying their “species“, branding each species with labels and giving additional information regarding their application area. • Aim: enhancing precision and expressiveness of the folksonomy by adding semantics. For query expansion along semantic relations, for enhanced navigation, as basis for semantic-oriented display. • Achieved by – conflation of multi-language tags, – summarization of synonyms, – distinction of homonyms, – establishment of semantic relations, – field-based tagging
  • 17. Synonym interrelation and distinction of homonyms.
  • 18. Hidden Semantic Relations in Folksonomies Source for image: http:// www.flickr.com
  • 19. Tag Gardening Activities: 4. Fertilizing • Combination of folksonomies and KOS during indexing and retrieval. • Aim: enhancing precision and recall and the expressiveness of the folksonomies by adding semantics, for query expansion during retrieval via semantic relations, for enhanced indexing functionalities, for enhanced navigation within the folksonomy, as basis for semantic-oriented displays. • Achieved by – semantic-oriented tag suggestions during indexing and retrieval ( tag suggestions not based on tag popularity to avoid “success breeds success-effect”) – establishment of semantic relations by mappings to KOS (after indexing)
  • 20. Tag Gardening Activities: Fertilizing Type 1 and 2 • Fertilizing Type 1 Fertilizing a folksonomy with semantic structures from other knowledge organization systems. • Fertilizing Type 2 Fertilizing a structured knowledge organization system (ontology, thesaurus) with user-generated terminologies.
  • 21. Tag Gardening in Practice Possibilities – Editing and deleting functionalities for tags on personal and document level. – Detecting and labeling semantic relations. – Use of power tags as candidates on document collection level. – Co-occurence computations as starting points. – Tools for personal tag gardening on personomy level. – Collaborative tag gardening in (small) communities. – Power users as chief gardeners in communities.
  • 22. Gardening support: Power tags and co- occurences as candidates and starting points Steps: • Detecting power tags on the resource level • Computing co-occurences of power tags with the whole tag collection • Result: distribution and power tags on co- occurence level Power relations as candidates for new folksonomy structuring approaches Data retrieved 2008-05-15 from del.icio.us
  • 23. Personal Tag Gardening • Our approach: TagCare (www.tagcare.org, currently under development) • „take care of your tags“ TagCare imports personal tags from different tagging platforms and allows to manage them centrally for better personal overview. Personal structured tag vocabulary for: • Personal infomation management (PIM) • Consistent tagging across platforms (functionalities under development). • Search in folksonomy-based systems (e.g. use pre- cooked personal synonym lists).
  • 24. Personal Tag Gardening Use of TagCare – first version • Adding, deleting and editing of tags • Hierarchical structuring of tags • Interrelating of synonyms • Labeling of otherwise related tags • Statistics on tag frequencies
  • 25. Tag Gardening Community Solution for small, closed communties and development of shared vocabularies: Community tagging plus single information architects. Information architect: Information architect: builds a structured thesaurus builds a structured thesaurus and enhances it with tags. and enhances it with tags. establishes structures between establishes structures between loose tags (emergent semantics). loose tags (emergent semantics). Community: Community: applies tags and applies tags and performs minor editing performs minor editing activities. activities.
  • 26. Collaborative Tag Gardening Challenges and open questions – Who may perform editing actions? Who is an authorized user? – Which tags are spam? – How can users collaborate, which aspects are determined collectively? – Can tagging guidelines be applied? – …
  • 27. Challenges and open questions • The three elements of folksonomies (resources – people – tags) constitute a three-dimensional knowledge space: domain – community – expressiveness. • Conflicts are caused, if all three elements should be considered to large extents simultaneously.
  • 28.
  • 29. Conclusions & Outlook • Folksonomies may be enriched with semantics to achieve a combination of vocabulary dynamics and structure. • Aim: improved information retrieval on personomy and collection level. • Manual and semi-automatic approaches may be combined. • Tag gardening may first be applied to small communities or small application domains. • Solutions for bigger communities are needed, „chief gardeners“ as first approach. • Future work: development of TagCare, analysis of types of semantic relations for folksonomy enrichment, tagging for support of ontology or thesaurus development.
  • 30. Thank You Best regards from Düsseldorf! Contact: katrin.weller@uni-duesseldorf.de isabella.peters@uni-duesseldorf.de http://www.phil-fak.uni-duesseldorf.de/infowiss/