Standardized infrastructure for information or knowledge sharing is required to make autonomous agents interdependent on each other for effective collaboration in a multi-agent system. Folksonomy has become very popular as an enabling technology to provide a common conceptualization of the data that agent systems use. However, there are problems on free-form tagging in folksonomy. Folksonomy is only concerned with a group of instances which are labeled with tags without a formal definition. No available tool provides a way to contextualize folksonomies with respect to users, communities, goals, tasks, and so on. There is no formal approach to classifying and sharing tags that reflect a user's mental model of information resources in terms of folksonomy. We present a novel approach to developing an agent environment for contextualizing folksonomies in a triadic context using Formal Concept Analysis. We conducted an experiment to build concept hieracrhies and contextualize folksonomies from tags of blogosphere.
展开▼
