A collaborative tagging system with formal concept analysis

Abstract

Tags can be used to annotate resources on the web. This enables users to share or browse the resources or retrieve them in future. Collaborative Tagging systems or folksonomies have the potential to become an integral part of Web 2.0. Formal Concept Analysis (FCA) is a powerful tool commonly used in Artificial Intelligence, Data Mining and with the Semantic Web. FCA has been used in online document and resource management systems. In this case the resources are treated as objects and tags as attributes. FCA groups these resources hierarchically in a lattice structure thereby providing multiple dimensions to information retrieval. Objects are grouped with a set of attributes common to all of them. These groups are called concepts and are the building blocks of FCA lattices. A system is discussed that models objects and their tags with Formal Concept Analysis. A user's query for objects with certain attributes can be mapped to a particular concept. The objects of this concept can be returned as results. Further related and relevant results can be provided by finding the concepts most similar to the result concept and returning their objects to the user as well. Thus, an information retrieval system can be implemented. Further automation can be investigated with machine learning or artificial intelligence techniques.