RichTags: A Social Semantic Tagging System

Social tagging systems allow users associating arbitrary keywords (or tags, or labels) to resources they want to save for future recall. Such saved items are called posts or bookmarks and usually constitute shared information in social tagging systems (although access control mechanisms might be applied as well). This means that users of a social tagging system can save and share their bookmarks with each other. The term social stresses the fact that much of the usefulness of the system relies on the data the users submit and share with each other. As a member of this category of tools, RichTags aims to overcome some weaknesses of the conventional social tagging systems (folksonomies) by utilizing Semantic Web technologies. The defining characteristic of the system is that the tags constitute an ontology of meaningful concepts, which is collectively managed by the users of the system. Hence, the approach is called social semantic tagging. It overcomes the polysemy, the synonymy, and the basic level variation problems encountered in the conventional systems. As well, it offers higher precision and recall. Current realisation of semantic tagging basically concerns an effort to automatically derive semantics out of folksonomies without affecting the mechanism of tagging applied in them. In contrast, RichTags’s approach for semantic tagging is a social process relied on the collective intelligence of the users instead of automation methods. The later means that the users collectively expand the tag vocabulary throughout the tagging task, while consistency mechanisms are applied to keep the vocabulary consistent during this expansion. The basic factor that differentiates RichTags from existing proposals for the enhancement of tags with meaning is that the primary mechanism relies on human collective intelligence and not on automation methods. However, this does not mean that the proposed automation techniques could not be combined with RichTags; contrariwise they could be very useful to speed up the production of the initial set of semantic tags in the vocabulary. Finally, RichTags is not limited to enriching the tags with meaning as current efforts primarily aim to; instead it utilizes this semantic information to improve the tagging and the exploration tasks of tagging systems