SCOT Model and Folksonomy

From Suk Hyung’s comment, he and I try to represent intuitively the SCOT ontology.
In his comment he use three elements for defining a folksonomy as the following:

F = (P, T, R, Y)
where, P : the set of People
T : the set of Words(i.e. “Tags”)
R : the set of Resources
Y \subset P \times T \times R : it represents “tagging”.

There are lots of efforts to define formal model of a folksonomy. For example, Andreas defines the term personomy and folksonomy as a formal concept in his paper(Bibsonomy: A Social Bookmark and Publication Sharing System).

  • See Definition 1, 2
  • See Definition 3
  • I think the basic elements for the folksonomy should be a combination of (user, tag, resource). But when we use folksonomy from del.icio.us, flickr etc a resource is linked by a unique namespace, and user and tag is not a single value. I think it might be hard to describe three elements together.
    Ok. what is a main purpose of the SCOT model? The SCOT model aims to represent a folksonomy model. A folksonomy consists of a set of users, a set of tags, and a set of resources, and the set of tag must have frequencies of the individual tags.

    To do so, the SCOT model should have a container for three elements. The Tagcloud class has two core properties (scot:tagcloudOf, scot:locatedIn) to describe user, site information and the Tag class has a number of properties to describe semantics of tags. A specific resource can be linked the unique namespace in conjuntion with the tagcloud’s namespace and tag.

    Based on James’ model, we try to make abstract relation model for SCOT, SIOC, FOAF, SKOS, and folksonomy. Basically SOCT may has entire elements of folksonomy and could be connected to other RDF vocabularies.

    SCOT Ontology Model

    Class relationship

    Basically user can create tags in a site. A tagcloud has a set of tags.
    So we can extract the following elements:

    (User, Tag, Site, Tagcloud)

    and also we can think of multiple users, sub community in a site. The elements can be added;

    (User, Tag, Site, Tagcloud, User group, Forum)

    We can draw the following diagram to represent the relationships among elements.

    Class relationships

  • Tagger: User, User group - foaf:Agent, foaf:Person, foaf:Group
  • Location of Tagcloud: Site, Blog, Wiki, Forum - sioc:Site, sioc:Forum
  • SCOT Ontology Model

    The core concepts of SCOT include Tagcloud and Tag. Tagcloud class represents a collective semantics emerged from the tagging activities of the people who are socially connected while the Tag class provides a semantically rich way of representing the concept of a tag and the relationship among tags in a given social space. One notable difference with other approaches is the implication of resources that are linked by the unique tag namespace of SCOT. Generally a tag exists in a resource and a tagcloud as a set of tags has a link to connect among the tag and the resource. We do not describe each piece of resource information in the SCOT Ontology, but we can connect a resource of individual tag by URI mechanism.

    The SCOT ontology is linked to the three dimensional relationships that are represented in SIOC, FOAF, and SKOS. We use the SIOC concepts to describe site information
    and relationships among site-resources as well as site-site and use the FOAF concepts to represent either a human or machine agent because a tag can be generated manually by a human user or automatically by a machine. Also we try to represent the relationships among users. This relationship has the two aspects: user-user and user-user group. When we are tagging and are using them, we assume these relationships. Finally, we use the SKOS to allow semantically relate a tag with another tag such as skos:broader and skos:narrower.

    SCOT Ontology Model

    The Folksonomy Triangle

    It is quite simple model to explain what is folksonomy.

    folksonomy triangle

    copyright to all biscuit’s photos (http://www.flickr.com/photos/allbiscuit/)

    Folksonomy is overrated?

    I think folksonomy might be overrated. The vision and principles of folksonomy are reasonable. The problem stems from the realization.

  • Tagging is a user-oriented cognitive activity, but folksonomy is only statistically weighted terms lists
  • There is no meanings and relationships among each terms
  • Sharing and Reusing data limited to host site
  • Folksonomy itself is NOT solution
  • what do we need?

  • Minimal Structure
  • Minimal Semantic
  • Unique local tag namespace
  • Welcome to SCOT blog!

    Here you’ll find development, updates and plans on tagging innovations.