Todays topic

22
Todays topic Social Tagging By Christoffer Hirsimaa

description

Todays topic. Social Tagging. By Christoffer Hirsimaa. Stop thinking, start tagging: Tag Semantics arise from Collaborative Verbosity. Christian Körner , Dominik Benz, Andreas Hotho, Markus Strohmaier, Gerd Stumme From WWW2010. Where do Semantics come from?. - PowerPoint PPT Presentation

Transcript of Todays topic

Page 1: Todays topic

Todays topicSocial Tagging

By Christoffer Hirsimaa

Page 2: Todays topic

Stop thinking, start tagging:Tag Semantics arise from Collaborative VerbosityChristian Körner, Dominik Benz, Andreas Hotho,Markus Strohmaier, Gerd Stumme From WWW2010

Page 3: Todays topic

Where do Semantics come from?

Semantically annotated content is the „fuel“ of the next generation World Wide Web – but where is the petrol station?

Expert-built expensive

Evidence for emergent semantics in Web2.0 data Built by the crowd!

Which factors influence emergence of semantics? Do certain users contribute more than others?

3

Page 4: Todays topic

OverviewEmergent Tag Semantics

Pragmatics of tagging

Semantic Implications of Tagging Pragmatics

Conclusions

4

Page 5: Todays topic

Emergent Tag Semantics tagging is a simple and

intuitive way to organize all kinds of resources

formal model: folksonomy F = (U, T, R, Y) Users U, Tags T, Resources R Tag assignments Y (UTR)

evidence of emergent semantics Tag similarity measures can

identify e.g. synonym tags (web2.0, web_two)

5

Page 6: Todays topic

Tag Similarity Measures: Tag Context Similarity Tag Context Similarity is a scalable and precise tag

similarity measure [Cattuto2008,Markines2009]: Describe each tag as a context vectorEach dimension of the vector space correspond to

another tag; entry denotes co-occurrence countCompute similar tags by cosine similarity

5 30 1 10 50design

software blog web programming…JAVA

Will be used as indicator of emergent semantics!

6 / 20

6

Page 7: Todays topic

= tag

Assessing the Quality of Tag Semantics

JCN(t,tsim) = 3.68TagCont(t,tsim) = 0.74

Folksonomy Tags = synset

WordNet Hierarchy

Mapping

Average JCN(t,tsim) over all tags t: „Quality of semantics“

7

Page 8: Todays topic

bev

alc nalc

beer wine

Tagging motivation Evidence of different ways HOW users tag (Tagging

Pragmatics) Broad distinction by tagging motivation [Strohmaier2009]:

donutsduff

margebeer

bart

barty

Duff-beer

„Categorizers“…- use a small controlled tag

vocabulary- goal: „ontology-like“

categorization by tags, for later browsing

- tags a replacement for folders

„Describers“…- tag „verbously“ with freely chosen

words- vocabulary not necessarily consistent

(synonyms, spelling variants, …)- goal: describe content, ease retrieval

8

Page 9: Todays topic

Tagging Pragmatics: Measures How to disinguish between two types of

taggers?

Vocabulary size:

Tag / Resource ratio:

Average # tags per post:

high

low

9

Page 10: Todays topic

Orphan ratio:

R(t): set of resources tagged by user u with tag t

high

low

Tagging Pragmatics: Measures

10

Page 11: Todays topic

Tagging pragmatics: Limitations of measures Real users: no „perfect“ Categorizers /

Describers, but „mixed“ behaviour

Possibly influenced by user interfaces / recommenders

Measures are correlated

But: independent of semantics; measures capture usage patterns

11

Page 12: Todays topic

Influence of Tagging Pragmatics on Emergent Semantics Idea: Can we learn the same (or even better) semantics from the

folksonomy induced by a subset of describers / categorizers?

Extreme Categorizers

Extreme Describers

Complete folksonomy

Subset of 30% categorizers

= user

12

Page 13: Todays topic

Experimental setup

1. Apply pragmatic measures vocab, trr, tpp, orphan to each user2. Systematically create „sub-folksonomies“ CFi / DFi by

subsequently adding i % of Categorizers / Describers (i = 1,2,…,25,30,…,100)

3. Compute similar tags based on each subset (TagContext Sim.)4. Assess (semantic) quality of similar tags by avg. JCN distance

TagCont(t,tsim)= …

JCN(t,tsim)= …

DF20CF5

13

Page 14: Todays topic

Dataset

From Social Bookmarking Site Delicious in 2006

Two filtering steps (to make measures more meaningful):

Restrict to top 10.000 tags FULL Keep only users with > 100 resources MIN100RES

dataset |T| |U| |R| |Y|ORIGINAL 2,454,546 667,128 18,782,13

2 140,333,7

14 FULL 10,000 511,348 14,567,46

5117,319,0

16MIN100RES

9,944 100,363 12,125,176

96,298,409

14 / 20

14

Page 15: Todays topic

Results – adding Describers (DFi)

15

Page 16: Todays topic

Results – adding Categorizers (CFi)16

Page 17: Todays topic

Summary & Conclusions Introduction of measures of users‘ tagging motivation

(Categorizers vs. Describers)

Evidence for causal link between tagging pragmatics (HOW people use tags) and tag semantics (WHAT tags mean)

„Mass matters“ for „wisdom of the crowd“, but composition of crowd makes a difference („Verbosity“ of describers in general better, but with a limitation)

Relevant for tag recommendation and ontology learning algorithms

17

Page 18: Todays topic

My thoughts and remarks

Confirmed deleting spammers is useful once again, but how useful?

Try to recursively combine the set of describers / categorizers

18

Page 19: Todays topic

Q&A and discussion!

19

Page 20: Todays topic

Thank you for your attention!

20

Page 21: Todays topic

21 / 20

Extras:

21

Page 22: Todays topic

References [Cattuto2008] Ciro Cattuto, Dominik Benz, Andreas Hotho,

Gerd Stumme: Semantic Grounding of Tag Relatedness in Social Bookmarking Systems. In: Proc. 7th Intl. Semantic Web Conference (2008), p. 615-631

[Markines2009] Benjamin Markines, Ciro Cattuto, Filippo Menczer, Dominik Benz, Andreas Hotho, Gerd Stumme: Evaluating Similarity Measures for Emergent Semantics of Social Tagging. In: Proc. 18th Intl. World Wide Web Conference (2009), p.641-641

[Strohmaier2009] Markus Strohmaier, Christian Körner, Roman Kern: Why do users tag? Detecting users‘ motivation for tagging in social tagging systems. Technical Report, Knowledge Management Institute – Graz University of Technology (2009)

22