Upload
freddy-limpens
View
1.143
Download
0
Embed Size (px)
DESCRIPTION
This paper presents our approach to collaborative and semiautomatedsemantic structuring of folksonomies. Tags freely provided by users of online communities are not semanticallylinked, and this hinders signicantly the potentials for browsing and exploring these data. We propose a sociotechnicalsystem combining automatic handlings of tags, using state of the art algorithm, and user friendly interfaces designed after a careful analysis of the usage of our targetcommunities. Much like folksonomies, our socio-technical system lets each user maintain his own view while still beneting from others contributions. As a complement to similar approaches, our approach supports conflicting point ofviews all along the life-cycle of semantically enriched folksonomies.
Citation preview
1
Helping online communities to semantically enrich folksonomies
WebScience 2010, Raleigh, NC, USA, 26-27 april 2010
Freddy Limpens, Fabien Gandon Edelweiss, INRIA Sophia Antipolis
{freddy.limpens, fabien.gandon}@inria.fr
Michel BuffaKewi, I3S, Université Nice-Sophia Antipolis
Edelweiss
2
Social tagging for bookmarking
http://delicious.com
3
… and the resulting
FOLKS - ONOMIES
4
newyork = new_york = nyc
Spelling variations of tags
5
Amibiguity of tags
… or in Texas ?
.. inFrance ?
6
How to turn folksonomies ...
...into comprehensible topic structures ?
?
pollution
Soil pollutions
has narrower
pollutant Energy
related related
7
… by collectingall user's expertiseinto the process
8
Our approach
Integrate usage-analysis for a tailored solution
Supporting diverging points of view
Automatic processings
+
Human expertise through user-friendly interfaces
9
Concrete scenario
Expertsproduce docs
+ tag
Archivistscentralize + tag
Public audienceread + tag
10
Supporting diverging points of view
car pollutionskos:related
11
car pollutionskos:related
John Paul
Supporting diverging points of view
agrees disagrees
12
car pollutionskos:related
John Paul
Supporting diverging points of view
hasApproved hasRejected
tagSemanticStatement (named graph)
ADDING TAGS
Automatic processing
User-centricstructuring
Detect conflicts
Globalstructuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment
life-cycle
14
pollution
pollutantpollution
pollutionpollutionpollutionpollution Soil pollutions
1. Comparing Tag labels with string edit distances
15
Evaluation of 30 edit distances
Combining the best metrics
Needs complement !
1. Comparing Tag labels with string edit distances
16
Fig. Markines et al. (2009)
Association via:
Users
tags
2. Analyzing the tri-partite structure of folksonomies
17
pollution
Soil pollutions
has narrower
pollutant
Environment
Energy
related related
related
Result: automatically suggested semantics
ADDING TAGS
Automatic processing
User-centricstructuring
Detect conflicts
Globalstructuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment
life-cycle
19
Embedding structuring tasks within everyday activity (searching e.g)
20
Embedding structuring tasks within everyday activity (searching e.g)
21
Capturing user's point of view
ADDING TAGS
Automatic processing
User-centricstructuring
Detect conflicts
Globalstructuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment
life-cycle
23
Conflict detection
environment pollution
narrower
broader
24
Conflict detection
environment pollution
narrower
broader
Using rules e.g:
IF num(narrower)/num(broader) ≥ cTHEN narrower winsELSE 'more generic' wins
25
Conflict detection
environment pollution
narrower
broader
related
related
broader narrower
more generic more generic
ADDING TAGS
Automatic processing
User-centricstructuring
Detect conflicts
Globalstructuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment
life-cycle
27
environment pollutionrelated
ReferentUser
Global structuring by Referent
hasApproved
ADDING TAGS
Automatic processing
User-centricstructuring
Detect conflicts
Globalstructuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment
life-cycle
29
Take away message (conclusion)
30
Help communities
structure their tags
What we do :
31
Our contributions:
Usages analysis
Automatic processing of tags
Tag structuring embedded in every-day tools
Supporting multi-points of view
32
Implementation & tests
• ADEME dataset (~10000 tags)
• Tag server
• Tag searching interface
33
Perspectives
• More automatic methods
• More ontological resources
• Other interfaces (tagging, global structuring)
• Test + Evaluation @ Ademe & Orange Labs
34
Thank you !
http://www-sop.inria.fr/members/Freddy.Limpens/
http://isicil.inria.fr