SlideShare ist ein Scribd-Unternehmen logo
1 von 19
Downloaden Sie, um offline zu lesen
Knowledge Management Institute




              Pragmatic Evaluation of Folksonomies

                         20th International World Wide Web Conference (WWW2011)
                                               Hyderabad, India


                  D. Helic, M. Strohmaier, C. Trattner, M. Muhr, K. Lerman


                                                  Markus Strohmaier
                                 Assistant Professor, Graz University of Technology, Austria
                                           Visiting Scientist, (XEROX) PARC, USA




 Markus Strohmaier                                         2011
                                                                                               1
Knowledge Management Institute




               Taxonomies: Categorization by Experts

                        Taxonomy: Usually produced and maintained by
                             few (e g dozens of) domain experts
                                 (e.g.                  experts.

                                 Alternative: Folk-generated taxonomies
                                            („Folksonomies“)
                                            ( F lk      i “)


                                   But how useful are such hierarchical
                                 structures? How can they be evaluated?




 Markus Strohmaier                                2011
                                                                          2
Knowledge Management Institute




                                   Outline of this talk

            1. Folksonomies
                  Construction and E l ti
                  C   t ti       d Evaluation


            2.
            2 Decentralized Search
                  J. Kleinberg‘s algorithm


            3. Pragmatic Evaluation Framework
                  Presentation and discussion


            4. Results & Findings

 Markus Strohmaier                              2011
                                                          3
Knowledge Management Institute




                                   Outline of this talk

            1. Folksonomies
                  Construction and E l ti
                  C   t ti       d Evaluation


            2.
            2 Decentralized Search
                  J. Kleinberg‘s algorithm


            3. Pragmatic Evaluation Framework
                  Presentation and discussion


            4. Results & Findings

 Markus Strohmaier                              2011
                                                          4
Knowledge Management Institute




                   Tagging: Social classification by users

                                                            Users label and categorize
                    Resources                             resources with concepts (tags)

                                                                               Tags

              Users
              U

        is a tuple V:= (U, T, R, Y) where
        • th th
             the three di j i t fi it sets U T R correspond t
                       disjoint, finite t U, T,           d to                       user 1

               –    a set of persons or users u ∈ U
               –    a set of tags t ∈ T and
               –    a set of resources or objects r ∈ R                      tag 1            res. 1
        •      Y ⊆ U ×T ×R, called set of tag assignments
                                                                        Tag similarity based on
                                                                         users and resources
 Markus Strohmaier                                 2011
                                                                                                       5
Knowledge Management Institute




                         Construction of Folksonomies
   From tag centrality to tag tag centrality:
   F    t      t lit t high generality:
                          t          lit
                             more abstract




                                 low tag centrality:
                                     more specific

                                                  Other existing folksonomy algorithms:
                                                     k-means, affinity propagation, …
                                         [Heyman and Garcia-Molina 2006]
 Markus Strohmaier                                     2011
                                                                                          6
Knowledge Management Institute



              Semantic Evaluation of Folksonomies
     Emerging Hierarchy
         g g          y                                    Expert Hierarchy
                                                             p            y
     (Emergent)                                            (Golden Standard)
     via e.g. hierarchical clustering                      WordNet: a lexical DB for English

                                                                                  computers

                                                Map-                                Synset Hierarchy
       Programming                              ping
                                                                    programming
                             distance d1 = 1                                                distance
                                                                                             d2 = 2
                          Python
                                                              Design
                                                                   g            languages
                                                                                   g g
                                                              patterns
 abs. difference |d1 - d2| a                   Semantic
 simple p y for the q
     p proxy           quality
                             y                 grounding                 j
                                                                         java               python
 of emergent semantics
 Markus Strohmaier                                  2011
                                                                                                     8
Knowledge Management Institute




                                   Outline of this talk

            1. Folksonomies
                  Construction and E l ti
                  C   t ti       d Evaluation


            2.
            2 Decentralized Search
                  J. Kleinberg‘s algorithm


            3. Pragmatic Evaluation Framework
                  Presentation and discussion


            4. Results & Findings

 Markus Strohmaier                              2011
                                                          9
Knowledge Management Institute



                                                                        Decentralized Search
                                                                                                                                             Idea: use folksonomies as
       Then, the performance of decentralized search
                 p                                                                                                                             background knowledge
                                                                                                                                                   g                g
       Background knowledge:                                                                                                                Shortest path to target
       depends on the suitability of folksonomies.
        (a tag hierarchy)

       In other words, we can evaluate the suitability of
       folksonomies for decentralized search through
       simulations.                                                                                                                             Folksonomy Folksonomy                     Folksonomy
                                                                                                                                                     1          ...                            n




                                                                                                                                                         shortest path found with
        A (tag-tag) network:                                                                                                                             local k
                                                                                                                                                         l   l knowledge pLK = 4
                                                                                                                                                                    l d

        Goal: Navigate from START to TARGET                                                                                                                                              Δ = pLK-pGK
        using local and background knowledge
        only
                                                                                   candidates                      start                                                                      target
                                                                                                                                                                 shortest path with
                                                                                                                                                                          p
                                                                                                                                                             global knowledge pGK = 3
      Markus Strohmaier                                                                               2011
J. Kleinberg. The small-world phenomenon: An algorithmic perspective. Proc. 32nd ACM Symposium on Theory of Computing, 2000. Also appears as Cornell Computer Science Technical Report 99-1776 (October 1999)   10
Knowledge Management Institute




                                   Outline of this talk

            1. Folksonomies
                  Construction and E l ti
                  C   t ti       d Evaluation


            2.
            2 Decentralized Search
                  J. Kleinberg‘s algorithm


            3. Pragmatic Evaluation Framework
                  Presentation and discussion


            4. Results & Findings

 Markus Strohmaier                              2011
                                                          11
Knowledge Management Institute




                        Pragmatic Evaluation Framework
            General idea:
            • Use the OUTPUT produced by folksonomy algorithms
              (hierachical structures) as INPUT (b k
              (hi    hi l t t        )          (background
                                                          d
              knowledge) for decentralized search.

                        Framework                          Instantiation
                                                            K-means, Aff.Prop.,
                        1. Generate n folksonomies       DegCentrality, CloCentrality
                                                           exploratory navigation
                        2. Model navigational task
                                                            success rate, stretch
                        3. Select evaluation metrics
                                                            decentralized search
                        4. Simulate navigation
                        4 Sim late na igation
                                                           comparative evaluation
                        5. Evaluate performance


 Markus Strohmaier                                2011
                                                                                        12
Knowledge Management Institute




                        Simulating Exploratory Navigation
                                                                                      Topically
                                                                                        related
                              START                              TARGET                    tags
             tags



       resources

                                                                                           Topically
                                                                                             related
                                       Random                                             resources
                                                                           Random
                                                                           R d
                                           start
                                                                           resource
      Usefulness of:                  page: e.g.
                                        landing
                                      page from
                                         search
                                         engine           We generate 100.000 search pairs
                                                          (start, target) for each dataset, and
Folksonomy F lk
F lk       Folksonomy    Folksonomy
                         F lk                             run simulations
     1          ...           n
  Markus Strohmaier                                2011
                                                                                                  13
Knowledge Management Institute




                                   Outline of this talk

            1. Folksonomies
                  Construction and E l ti
                  C   t ti       d Evaluation


            2.
            2 Decentralized Search
                  J. Kleinberg‘s algorithm


            3. Pragmatic Evaluation Framework
                  Presentation and discussion


            4. Results & Findings

 Markus Strohmaier                              2011
                                                          14
Knowledge Management Institute




   Success Rates Across Different Folksonomies
                                 flickr dataset
                                                                Tag generality
                                                                approaches
                                                                k-means /
                                                                affinity propagation


                                                                Random
                                                                folksonomy
                  Success rate:
                  The number of times an agent is successful
                  in finding a path using a particular
                  folksonomy as background knowledge           All approaches outperform a
                                                               random folksonomy y
                                              n
           max hops n: the maximal number of steps an agent
                                                               Tag generality approaches
           is allowed to perform before stopping (a tunable outperform k-means / Aff.
           parameter e.g., an agent only f ll
                   t               t l follows n li k )
                                                   links).  Propagation
 Markus Strohmaier                                2011
                                                                                             16
Knowledge Management Institute




             Success Rates Across Different Datasets




 Holds for all                                    But how
  datasets                                      efficient are
   (to diff.
       diff                                         those
  extents)                                     folksonomies
                                                   during
                                                  search?

 Markus Strohmaier               2011
                                                          17
Knowledge Management Institute



                                 Stretch Δ = pLK-pGK
                                                 p
                 Shortest Paths found with Local Knowledge
                                     Bibsonomy K M
                                     Bib       K-Means


                                            Finds no path:
                                            Δ = infinite
                                            Finds paths that is +1 longer:
                                            Δ=1
 Holds for all
  datasets
  d t      t                                Finds shortest possible path:      Tag
                                                                               T generality
                                                                                          lit
   (to diff.                                Δ=0                              approaches (d+e)
  extents)                                                                   find much shorter
                                                                                   paths!




 Markus Strohmaier                       2011
                                                                                                 18
Knowledge Management Institute




                        Pragmatic Evaluation Framework

                      Framework                           Instantiation            Alternatives
                                                          K-means, Aff.Prop.,       other folksonomy
                      1. Generate n folksonomies            DegCentrality,            algorithms or
                                                             CloCentrality          expert taxonomies
                                                              exploratory               other tasks
                      2. Model navigational task              navigation
                                                          success rate, stretch   other evaluation metrics
                      3.
                      3 Select evaluation metrics
                                                          decentralized search       actual click data
                      4. Simulate navigation
                                                              comparative            other evaluation
                      5. Evaluate performance                  evaluation              approaches




                     Pragmatic evaluation produces different results for different
                     tasks and different assumed or observed navigation behavior.

                     The evaluation framework is completely general with regard to
                     the task, data and evaluation metrics adopted.
 Markus Strohmaier                                 2011
                                                                                                             19
Knowledge Management Institute




                            Results & Findings: Summary
            1. Folksonomies are useful b k
            1 F lk         i       f l background k
                                                d knowledge f
                                                      l d for
               navigation.

            2. Existing folksonomy algorithms are more useful
               than a random baseline.
                               baseline

            3.
            3 Tag generality approaches perform better than
               existing hierarchical clustering approaches.

            4. Pragmatic results support theoretical analysis
               (not presented in talk – included in paper).
 Markus Strohmaier                     2011
                                                                20
Knowledge Management Institute




                                            Thank You.
                                            Th k Y

                                           Markus Strohmaier
                                       markus.strohmaier@tugraz.at



                                 D. Helic, M. Strohmaier, C. Trattner, M. Muhr, K. Lerman
                                           Pragmatic Evaluation of Folksonomies
                               20th International World Wide Web Conference (WWW2011)
                                     Hyderabad, India, March 28 - April 1, ACM, 2011.
                    http://kmi.tugraz.at/staff/markus/documents/2011_WWW2011_Folksonomies.pdf


 Markus Strohmaier                                 2011
                                                                                                21

Weitere ähnliche Inhalte

Andere mochten auch (7)

Life Spiel
Life SpielLife Spiel
Life Spiel
 
Star Wars in Second Life
Star Wars in Second LifeStar Wars in Second Life
Star Wars in Second Life
 
Life Spiel
Life SpielLife Spiel
Life Spiel
 
Extracting semantics from crowds
Extracting semantics from crowdsExtracting semantics from crowds
Extracting semantics from crowds
 
Star Wars in Second Life
Star Wars in Second LifeStar Wars in Second Life
Star Wars in Second Life
 
Rebekahs slideshow
Rebekahs slideshowRebekahs slideshow
Rebekahs slideshow
 
Rebekah Moton
Rebekah MotonRebekah Moton
Rebekah Moton
 

Ähnlich wie Pragmatic evaluation of folksonomies

Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Antonella Dattolo
 
Tag And Tag Based Recommender
Tag And Tag Based RecommenderTag And Tag Based Recommender
Tag And Tag Based Recommender
gu wendong
 
IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...
IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...
IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...
IMPACT Centre of Competence
 

Ähnlich wie Pragmatic evaluation of folksonomies (20)

Extracting Semantics from Crowds
Extracting Semantics from CrowdsExtracting Semantics from Crowds
Extracting Semantics from Crowds
 
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
 
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
 
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
Visualizing and Managing Folksonomies, SASWeb 2011 workshop, at UMAP 2011
 
Tags as tools for social classification
Tags as tools for social classificationTags as tools for social classification
Tags as tools for social classification
 
The Role Of Ontology In Modern Expert Systems Dallas 2008
The Role Of Ontology In Modern Expert Systems   Dallas   2008The Role Of Ontology In Modern Expert Systems   Dallas   2008
The Role Of Ontology In Modern Expert Systems Dallas 2008
 
Di d poster_dlf.pptx
Di d poster_dlf.pptxDi d poster_dlf.pptx
Di d poster_dlf.pptx
 
Of Categorizers and Describers: An Evaluation of Quantitative Measures for Ta...
Of Categorizers and Describers: An Evaluation of Quantitative Measures for Ta...Of Categorizers and Describers: An Evaluation of Quantitative Measures for Ta...
Of Categorizers and Describers: An Evaluation of Quantitative Measures for Ta...
 
A SVM Applied Text Categorization of Academia-Industry Collaborative Research...
A SVM Applied Text Categorization of Academia-Industry Collaborative Research...A SVM Applied Text Categorization of Academia-Industry Collaborative Research...
A SVM Applied Text Categorization of Academia-Industry Collaborative Research...
 
2012 04-26-ifip-wg.pptx
2012 04-26-ifip-wg.pptx2012 04-26-ifip-wg.pptx
2012 04-26-ifip-wg.pptx
 
Tag And Tag Based Recommender
Tag And Tag Based RecommenderTag And Tag Based Recommender
Tag And Tag Based Recommender
 
On the Navigability of Social Tagging Systems
On the Navigability of Social Tagging SystemsOn the Navigability of Social Tagging Systems
On the Navigability of Social Tagging Systems
 
Model-Driven Research in Social Computing
Model-Driven Research in Social ComputingModel-Driven Research in Social Computing
Model-Driven Research in Social Computing
 
Generating domain specific sentiment lexicons using the Web Directory
Generating domain specific sentiment lexicons using the Web Directory Generating domain specific sentiment lexicons using the Web Directory
Generating domain specific sentiment lexicons using the Web Directory
 
LAK13 linkedup tutorial_evaluation_framework
LAK13 linkedup tutorial_evaluation_frameworkLAK13 linkedup tutorial_evaluation_framework
LAK13 linkedup tutorial_evaluation_framework
 
Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...
Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...
Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...
 
Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...
Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...
Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neur...
 
IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...
IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...
IMPACT Final Event 26-06-2012 - Franciska de Jong - Indexing and searching of...
 
Pattern-based competence management
Pattern-based competence managementPattern-based competence management
Pattern-based competence management
 
Ontology Mapping
Ontology MappingOntology Mapping
Ontology Mapping
 

Kürzlich hochgeladen

Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Joaquim Jorge
 
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
?#DUbAI#??##{{(☎️+971_581248768%)**%*]'#abortion pills for sale in dubai@
 

Kürzlich hochgeladen (20)

From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
 
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
 
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
 
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
HTML Injection Attacks: Impact and Mitigation Strategies
HTML Injection Attacks: Impact and Mitigation StrategiesHTML Injection Attacks: Impact and Mitigation Strategies
HTML Injection Attacks: Impact and Mitigation Strategies
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdfUnderstanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024
 
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 

Pragmatic evaluation of folksonomies

  • 1. Knowledge Management Institute Pragmatic Evaluation of Folksonomies 20th International World Wide Web Conference (WWW2011) Hyderabad, India D. Helic, M. Strohmaier, C. Trattner, M. Muhr, K. Lerman Markus Strohmaier Assistant Professor, Graz University of Technology, Austria Visiting Scientist, (XEROX) PARC, USA Markus Strohmaier 2011 1
  • 2. Knowledge Management Institute Taxonomies: Categorization by Experts Taxonomy: Usually produced and maintained by few (e g dozens of) domain experts (e.g. experts. Alternative: Folk-generated taxonomies („Folksonomies“) ( F lk i “) But how useful are such hierarchical structures? How can they be evaluated? Markus Strohmaier 2011 2
  • 3. Knowledge Management Institute Outline of this talk 1. Folksonomies Construction and E l ti C t ti d Evaluation 2. 2 Decentralized Search J. Kleinberg‘s algorithm 3. Pragmatic Evaluation Framework Presentation and discussion 4. Results & Findings Markus Strohmaier 2011 3
  • 4. Knowledge Management Institute Outline of this talk 1. Folksonomies Construction and E l ti C t ti d Evaluation 2. 2 Decentralized Search J. Kleinberg‘s algorithm 3. Pragmatic Evaluation Framework Presentation and discussion 4. Results & Findings Markus Strohmaier 2011 4
  • 5. Knowledge Management Institute Tagging: Social classification by users Users label and categorize Resources resources with concepts (tags) Tags Users U is a tuple V:= (U, T, R, Y) where • th th the three di j i t fi it sets U T R correspond t disjoint, finite t U, T, d to user 1 – a set of persons or users u ∈ U – a set of tags t ∈ T and – a set of resources or objects r ∈ R tag 1 res. 1 • Y ⊆ U ×T ×R, called set of tag assignments Tag similarity based on users and resources Markus Strohmaier 2011 5
  • 6. Knowledge Management Institute Construction of Folksonomies From tag centrality to tag tag centrality: F t t lit t high generality: t lit more abstract low tag centrality: more specific Other existing folksonomy algorithms: k-means, affinity propagation, … [Heyman and Garcia-Molina 2006] Markus Strohmaier 2011 6
  • 7. Knowledge Management Institute Semantic Evaluation of Folksonomies Emerging Hierarchy g g y Expert Hierarchy p y (Emergent) (Golden Standard) via e.g. hierarchical clustering WordNet: a lexical DB for English computers Map- Synset Hierarchy Programming ping programming distance d1 = 1 distance d2 = 2 Python Design g languages g g patterns abs. difference |d1 - d2| a Semantic simple p y for the q p proxy quality y grounding j java python of emergent semantics Markus Strohmaier 2011 8
  • 8. Knowledge Management Institute Outline of this talk 1. Folksonomies Construction and E l ti C t ti d Evaluation 2. 2 Decentralized Search J. Kleinberg‘s algorithm 3. Pragmatic Evaluation Framework Presentation and discussion 4. Results & Findings Markus Strohmaier 2011 9
  • 9. Knowledge Management Institute Decentralized Search Idea: use folksonomies as Then, the performance of decentralized search p background knowledge g g Background knowledge: Shortest path to target depends on the suitability of folksonomies. (a tag hierarchy) In other words, we can evaluate the suitability of folksonomies for decentralized search through simulations. Folksonomy Folksonomy Folksonomy 1 ... n shortest path found with A (tag-tag) network: local k l l knowledge pLK = 4 l d Goal: Navigate from START to TARGET Δ = pLK-pGK using local and background knowledge only candidates start target shortest path with p global knowledge pGK = 3 Markus Strohmaier 2011 J. Kleinberg. The small-world phenomenon: An algorithmic perspective. Proc. 32nd ACM Symposium on Theory of Computing, 2000. Also appears as Cornell Computer Science Technical Report 99-1776 (October 1999) 10
  • 10. Knowledge Management Institute Outline of this talk 1. Folksonomies Construction and E l ti C t ti d Evaluation 2. 2 Decentralized Search J. Kleinberg‘s algorithm 3. Pragmatic Evaluation Framework Presentation and discussion 4. Results & Findings Markus Strohmaier 2011 11
  • 11. Knowledge Management Institute Pragmatic Evaluation Framework General idea: • Use the OUTPUT produced by folksonomy algorithms (hierachical structures) as INPUT (b k (hi hi l t t ) (background d knowledge) for decentralized search. Framework Instantiation K-means, Aff.Prop., 1. Generate n folksonomies DegCentrality, CloCentrality exploratory navigation 2. Model navigational task success rate, stretch 3. Select evaluation metrics decentralized search 4. Simulate navigation 4 Sim late na igation comparative evaluation 5. Evaluate performance Markus Strohmaier 2011 12
  • 12. Knowledge Management Institute Simulating Exploratory Navigation Topically related START TARGET tags tags resources Topically related Random resources Random R d start resource Usefulness of: page: e.g. landing page from search engine We generate 100.000 search pairs (start, target) for each dataset, and Folksonomy F lk F lk Folksonomy Folksonomy F lk run simulations 1 ... n Markus Strohmaier 2011 13
  • 13. Knowledge Management Institute Outline of this talk 1. Folksonomies Construction and E l ti C t ti d Evaluation 2. 2 Decentralized Search J. Kleinberg‘s algorithm 3. Pragmatic Evaluation Framework Presentation and discussion 4. Results & Findings Markus Strohmaier 2011 14
  • 14. Knowledge Management Institute Success Rates Across Different Folksonomies flickr dataset Tag generality approaches k-means / affinity propagation Random folksonomy Success rate: The number of times an agent is successful in finding a path using a particular folksonomy as background knowledge All approaches outperform a random folksonomy y n max hops n: the maximal number of steps an agent Tag generality approaches is allowed to perform before stopping (a tunable outperform k-means / Aff. parameter e.g., an agent only f ll t t l follows n li k ) links). Propagation Markus Strohmaier 2011 16
  • 15. Knowledge Management Institute Success Rates Across Different Datasets Holds for all But how datasets efficient are (to diff. diff those extents) folksonomies during search? Markus Strohmaier 2011 17
  • 16. Knowledge Management Institute Stretch Δ = pLK-pGK p Shortest Paths found with Local Knowledge Bibsonomy K M Bib K-Means Finds no path: Δ = infinite Finds paths that is +1 longer: Δ=1 Holds for all datasets d t t Finds shortest possible path: Tag T generality lit (to diff. Δ=0 approaches (d+e) extents) find much shorter paths! Markus Strohmaier 2011 18
  • 17. Knowledge Management Institute Pragmatic Evaluation Framework Framework Instantiation Alternatives K-means, Aff.Prop., other folksonomy 1. Generate n folksonomies DegCentrality, algorithms or CloCentrality expert taxonomies exploratory other tasks 2. Model navigational task navigation success rate, stretch other evaluation metrics 3. 3 Select evaluation metrics decentralized search actual click data 4. Simulate navigation comparative other evaluation 5. Evaluate performance evaluation approaches Pragmatic evaluation produces different results for different tasks and different assumed or observed navigation behavior. The evaluation framework is completely general with regard to the task, data and evaluation metrics adopted. Markus Strohmaier 2011 19
  • 18. Knowledge Management Institute Results & Findings: Summary 1. Folksonomies are useful b k 1 F lk i f l background k d knowledge f l d for navigation. 2. Existing folksonomy algorithms are more useful than a random baseline. baseline 3. 3 Tag generality approaches perform better than existing hierarchical clustering approaches. 4. Pragmatic results support theoretical analysis (not presented in talk – included in paper). Markus Strohmaier 2011 20
  • 19. Knowledge Management Institute Thank You. Th k Y Markus Strohmaier markus.strohmaier@tugraz.at D. Helic, M. Strohmaier, C. Trattner, M. Muhr, K. Lerman Pragmatic Evaluation of Folksonomies 20th International World Wide Web Conference (WWW2011) Hyderabad, India, March 28 - April 1, ACM, 2011. http://kmi.tugraz.at/staff/markus/documents/2011_WWW2011_Folksonomies.pdf Markus Strohmaier 2011 21