Search using Metadata, Semantic, and Ontologies

Traditional search techniques establish a direct connection between the information provided by users with the search engine. Users are only allowed to specify a set of keywords that will be syntactically matched against a database of keywords and references. This simple approach has several drawbacks since it gives rise to a low precision (the ratio of positive results with respect to the total number of false and positive results retrieved) and low recall (the ratio of positive results retrieved with respect to the total number of positive results in the reference base). Many factors influence this low precision and recall, namely polysemy and synonymy. In the first case, one word specified in a query might have several meanings and, in the second case, distinct words may designate the same concept. If appropriate strategies are used and included in a new generation of search engines, the number of false results can be drastically reduced. As a result, the impact of these two degrading factors can be reduced and even eliminated. As the interconnection of research areas such as artificial intelligence, semantic web, and linguistics becomes stronger and more mature, it is reasonable to explore how better search engines can be developed to more adequately respond to users’ needs. A new kind of search engine that has been explored for a few years now has been termed “semantic-based search engines” by many researchers. The underlying paradigm of these engines is to find resources based on similar concepts and logical relationships and not just similar words. These engines typically rely on the use of metadata, controlled vocabularies, thesauri, taxonomy, and ontologies to describe the searchable resources to ensure that the most relevant items of information are returned. The intend of this special issue is to bring together a compilation of recent research and developments toward the creation of a new paradigm for search engines that relies on metadata, semantics and ontologies, by providing readers with a “broad spectrum vision” of the most important issues on semantic search engines. One of the main problems concerns the recognition of items of interest in web documents