Download - Tagging Systems
Tagging SystemsAustin Wester
Tags A keywords linked to a resource
(image, video, web page, blog, etc) by users without using a controlled vocabulary.
They help to improve search, personal organization, metadata, spam detection, and reputation systems
Tag Purposes Social bookmarking
Personal bookmarks Store and retrieve resources
Social tagging systems Shared tags for particular resources Each tag is a link to additional resources
tagged the same way by other users
Examples of Tagging Systems Flickr: A site for sharing and viewing photos. It allows
users to store and tag their personal photos, tag friends photos and create a contact list
Del.icio.us: A “social bookmarking site.” It allows users to tag web pages for easy retrieval.
CiteULike: This site allows users to tag citations and references, e.g. academic papers or books.
Youtube: A collection of videos allowing users to view and share by placing tags on the videos.
Last.fm: A music information database allowing members to tag artists, albums, and songs
A model
Issue with Vocabulary Users use different terms to describe the
same resources Polysemy: A single word has multiple
meanings Synonymy: Different words have the same
meaning Abstraction: Resource can be tagged at
different levels of abstraction Animal, dog, German Shepherd, Alsatian
Different languages Missing context: Tags that have no real
relation with the images Holiday, me, friends, a person’s name
Taxonomy of Tagging Systems System design and attributes
How the characteristics of a tagging system effects the content, the tags and the usage
Users How their incentives and motivations
affect the tagging system
System Design and Attributes Tagging rights: A tag can be added or removed by the
creator of the resource, a restricted group or everyone Tagging support: The mechanism of a tag entry
Blind tagging: a user cannot see a resource’s tags added by others
Viewable tagging: all tags are visible Suggestive tagging: the system suggests possible tags to the
user Aggregation: System either allows duplicate tagging (bag-
model) or they prevent it (set-model) Type of object: images, videos, songs, web pages, blogs,
games, etc Source of Material: Resources that can be tagged can be
anything on the web, provided by users or by the system Resource connectivity: links, groups etc. connecting
resources other than tags Social connectivity: The connection between the users may
result in localized folksonomies.
User Incentives and Motivations Future retrieval: To mark individual or a
collection of resource items for later personal retrieval
Contribution and sharing: To add to conceptual clusters for the use by others of either a known or unknown audience
Attract attention: to draw others to their resources (common tags, spam tags)
Self presentation: to leave a mark Express opinions: to share their opinions with
others
My Research Flickr.com
Image popularity vs. tags Is there any relation
Flickr API
Related Work
Case Study: Flickr
Flickr is a photo-sharing site that considers tags as a core element to the sharing, retrieval, navigation and discovery of user-contributed images.
It allows users to upload their photos and share with the public.
Users can tag, choose favorites, comment, join groups, send messages, create networks, explore etc.
It contains user-contributed resources instead of global resources.
It allows self-tagging instead of free-for-all tagging. The tags are aggregated in sets instead of bags. It affords blind-tagging instead of suggested-tagging This system design motivates people to tag.
By Yahoo! Research Berkeley & UC Berkley School of Information
Flickr Tags Tags are not mandatory User can tag their friends’ photos.
But within 58 million tags observed, the overwhelming majority are owner tags.
A large group of people have very few distinct tags while a small group has extremely large sets of tags.
Tag vocabulary sizeacross the set of users
Usefulness and importance of tags pair-wise Pearson correlation between
The number of uploaded photos The count of user’s distinct tags The number of contacts designated by
the user
Flickr usage correlation
Growth of distinct tags 10 random users were chosen
Frequent uploaders ( > 100 photos) Frequent taggers ( > 100 tags)
The number of distinct tags were observed as the number of photos uploaded increased.
Growth of distinct tags
Vocabulary Formation Flickr allows social networks and
interest groups. There is a huge potential for social
influence in the development of tag vocabularies.
People can follow updates from their contacts and this promotes constant tagging.
Tag Categories
Number of tags per photo in Flickr
Screenshot of the river metaphor.By Yahoo! Research
Shows interesting tags during the current time period
http://research.yahoo.com/taglines/
Conclusion Social tagging systems have the potential to
improve many information systems problems.
Tagging systems could be improved Preventing problems of meaning Finding relations between the tags (synonyms,
abstractions) My research will be to see if there is any
relationship between the popularity of an image and the tags used to describe it
QUESTIONS?