Contextual Semantic Search Navigation
MetadataShow full item record
Search has for a long time been an important tool for users seeking information. Many popular search systems have been based on keyword search. However, such systems are not that well adapted to explore the search space. Effective exploration requires systems which can guide the user in his quest. The main focus of this thesis is to generate a framework for semantic navigational search which is compatible with the vector space model and unites formal and informal semantics. In this thesis, formal semantics is represented by ontologies, while informal semantics is represented by folksonomies. The research conducted has been divided into four main phases with different foci that together fulfill our research focus: (i) Formal semantics, (ii) Informal semantics, (iii) Structure verification, and (iv) Contextual semantic search navigation. In the first phase, the focus was ontologies and extending concepts with concept signatures. A concept signature is a shallow semantic representation of a concept based on how the concept is referred to in textual documents bridging the semantic and syntactic representation of concepts. The second phase focused on structuring folksonomy tags and extending tags with concept signatures. This phase also focused on consolidating the tag vocabulary. These two phases together unify formal and informal semantics with a common format. The third phase developed an approach for assessing the hierarchical relations of ontologies and structured folksonomies based on concept signatures. Finally, the fourth research phase developed an approach to contextual semantic search navigation based on a hierarchical model and concept signatures. The approach is based on semantic navigators which use the concept signatures actively to generate and rank navigation suggestions to the user. Several experiments have been conducted to evaluate the proposed approaches, and they show promising results. The main contributions of this thesis can be summarized as follows: • A semantic representation which unifies formal and informal semantics. • An unsupervised approach to identifying semantic similarity in tags. • An approach to contextualizing semantic navigators for search using tag signatures and hierarchical relations. • An approach to evaluating the taxonomic quality of ontologies and folksonomies. • An approach to generating concept signatures from formal semantics (ontology). • An approach to generating concept signatures from informal semantics (folksonomy).