Social Tagging

Technologies: Search for Meaning: Collaborative Tagging, Emergent Ontologies

UofS Social Computing Presenter: Mayya Sharipova

Papers y Shirky C. (2005) Ontology is Overrated y Marlow C., Naaman M., Boyd D., Davis M. (2006) HT06,

tagging paper, taxonomy, Flickr, academic article, to read

Ontology. Classification y A lot of information-> Organization of information

Ontology Study of entities and their relations

Classification, categorization Organization of a collection of entities into related groups

Categorization (examples) Dewey Decimal System's categorization Dewey, 200: Religion 210 Natural theology 220 Bible 230 Christian theology 240 Christian moral & devotional theology 250 Christian orders & local church 260 Christian social theology 270 Christian church history 280 Christian sects & denominations 290 Other religions

Library of Congress D: History (general) DA: DB: DC: DD: DE: DF: DG: DH: DJ:

Great Britain Austria France Germany Mediterranean Greece Italy Low Countries Netherlands

Yahoo directory

DK: DL: DP: DQ: DR: DS: DT: DU: DX:

Former Soviet Union Scandinavia Iberian Peninsula Switzerland Balkan Peninsula Asia Africa Oceania Gypsies

Problems with categorization/ontology y A lot of human labour y World of user ≠ world of experts y Constrained vocabulary y Binary decision y Subjective y Doesn’t fit changing world

Solutions y Links structure y Google search engine

y Social tagging y del.icio.us

Social tagging y Web-based tagging systems allow participants to annotate a

particular resource, such as a web page, a blog post, an image, a physical location etc. with a freely chosen set of keywords (“tags”). y Folksonomy instead of taxonomy y Del.icio.us, Technorati, Upcoming and Flickr

Advantages/ attributes of social tagging y Distributed workload of creating metadata y Market logic y Individual motivation, group value y Individual differences don’t have to be homogenized y User and time are core attributes y No signal lose y Filtering is done post hoc y Merged from URLs, not categories y Merges are probabilistic, not binary; no need for shelf

A model of tagging system

Taxonomy of tagging systems System design

User incentives influence resultant tags

System design

Tagging rights

Tagging support

Aggregation model

Resource connectivity

implications

Object type

Social connectivity

?

resultant tags

Source of material

Upcoming - events

Flickr - photos

Del.icio.us - bookmarks

User incentives Organization need (personal)

Future retrieval

Social needs (communicative) Attract Play & Contribution attention competition & sharing Self Opinion presentation expression

implications

?

resultant tags

Flickr study

How social tags can be used? y Better search engines y Information organization y Discovery & communication y Spam filtering y Reducing effect of link spam y Identifying trends and emerging topic globally

and within communities y Locating experts and opinion leaders in specific domain