Browsing by Author "Keddari, Djalila"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Using tags Associated to Resources to Build Ontologies
    (2014-08-11) Keddari, Djalila; Mellah, Hakima; Benna, Amel
    Social tagging has recently emerged in collaborative web (Web 2.0) as a support shared resources organization allowing users to categorize these resources (web pages, video, images) by associating them with keywords, called tags. However, the use of uncontrolled tags poses several problems, namely ambiguity, writing variations (many tags denoting the same concept) as well as tags volume. The latter has led to tags classification into several clusters using the K-Means algorithm. The purpose of this work is that from tags clusters and their hierarchical classification, it would be important to build an ontology as these tags have a semantic aspect and are expressed by users who used the tagged resources. Building ontologies based on resource tag is a way to make automatic their building without requiring experts in their building as usual. Reproducing these tags clusters into ontological form is a way to share its knowledge with other users. We aim through this paper to build an ontology based on tags. A technical approach to form ontological portions, starting from tags clusters is based on semantic distance existing between tags. As tags clusters have a semantic relationship, the semantic distance between clusters is used to merge ontological portions into a global ontology
  • Thumbnail Image
    Item
    Using tags Associated to Resources to Build Ontologies
    (CERIST, 2014-06-18) Keddari, Djalila; Mellah, Hakima; Benna, Amel
    Social tagging has recently emerged in the collaborative web as a support shared resources organization allowing users to categorize these resources, that can be web pages, video, or images, by associating them with keywords, called tags. However, the use of uncontrolled tags poses several problems, namely ambiguity, writing variations where many tags denote the same concept, as well as tags volume. The latter has led to tags classification into several clusters using the K Means algorithm. The purpose of this work is that from tags clusters and their hierarchical classification, it would be important to build an ontology as these tags have a semantic aspect and are expressed by users who used the tagged resources. Building ontologies, based on resources tags, is a way to make automatic their building without requiring experts as usual. Reproducing these tags clusters into ontological form is a way to share its knowledge with other users. We aim through this paper to build an ontology based on tags. A technical approach to form ontological portions, starting from tags clusters is based on semantic distance existing between tags. As tags clusters have a semantic relationship, the semantic distance between clusters is used to merge ontological portions into a global ontology.