2.1. Thesaurus
A thesaurus is the most widely used index structure in conventional information retrieval systems. It consists of a set of concepts and a limited set of relationships between these concepts. Three types of interconcept relationship are represented: equivalence (preferred/nonpreferred equivalent concepts), hierarchical (broader/narrower concepts), associative (related concepts).
As a result, the thesaurus consists of a standardized, controlled vocabulary of concepts that are hierarchically structured into a single inheritance tree. The major advantages of a thesaurus-based index structure are its flexibility and its intelligibility. The major drawback is that a lot of efforts are needed for thesaurus construction and validation. Although tools and techniques have been developed for the computer-assisted creation of' thesauri, most existing thesauri have been carefully handcrafted (e.g.. the well-known MeSH thesaurus). There are just a few applications of thesauri to index hypermedia information.
2.2. Faceted Thesaurus
Faceted thesaurus is an indexing technique in which concepts are classified into separate hierarchical structures and in which each hierarchy captures a different viewpoint of the documents, as in example presented in figure 5. A faceted thesaurus consists of a number of different thesauri, and each thesaurus is used to index the documents with respect to some different knowledge domain. The main advantage of using faceted thesaurus is that this index structure allows for greater exhaustivity and precision in the hypermedia indexing process. The disadvantages are the same as for those of conventional thesaurus, that is the creation of the thesauri and the validation of their hierarchical structure is very time-consuming, and these procedures are difficult to make computer-aided ones.
Fig. 5. The schematics of faceted thesaurus
2.3. Concept Lattice
A concept lattice is an extension (a powerful one) of the thesaurus index structure. This structure can be described mathematically a concept lattice is a partially ordered set of concepts in which every pair of concepts has both a greatest lower bound (a unique narrower concept) and a least upper bound (a unique broader concept), as presented in figure 6.
Fig. 6. The schema of a concept lattice for the concept of beverages
The notion of a concept lattice is very common in AI, in so-called ontologies. A principle approach called formal concept analysis provides a conceptualization called a concept lattice, and ontology is a specification of a concept lattice. In practice, ontology specifies the named part of a concept lattice called a concept space. Based upon the ontological relations, concept lattices and concept spaces are faceted, with each facet representing a distinct dimension of information. Then formal concepts with the subtype ordering form a class hierarchy, together with the named class of attributes. Every pair (in fact, any subcollection) of concepts has an associated meet concept corresponding to logical and an associated join concept corresponding to logical or. Tree hierarchies, for example file system hierarchies, are very special cases of lattices; they need just a bottom node to be added.
The major advantage of a concept lattice is that it can represent more flexible hierarchical structures than an ordinary or faceted thesaurus. There also exists a complete set of mathematical techniques that can be used to create concept lattices and check their internal consistency.
An example of information system, using concept lattice is given by the WorldView system [1], which is designed to process electronic news, articles, and abstracts of technical reports. Documents are automatically indexed and classified with respect to a lattice of concepts derived from the IEEE Inspec thesaurus. The WorldView retrieval engine interprets a user's query relative to this lattice of concepts, and then restricts the lattice to the sublattice relevant to the query. Using this sublattice, it can find the narrower concepts that can be used to extend the scope of the original query.
Quelle: http://www.maf.vu.lt/katedros/cs2/ajrb.pdf
