Abstract

The internet has facilitated the creation of large collections of documents by the aggregation of documents sent from different suppliers. Europeana is an example. In general, these documents have been identified using different thesauri and, at the same time, different languages. There are no systems that help the user to consult these documents using natural language, and also in their own language. Europeana has set up a prototype of a semantic search engine. In this text we analyze its structure and functioning. The conclusion, however, is that this is a very experimental version.

Abstract

With the increasing number of web sites, the time spent by users reviewing the results also increases. In addition, the nature of web content is semantically heterogeneous and oriented to people who will be able to understand it. Frequently the results from search engines do not correspond to the expected topic. One approach to improve the results is to match the content of the web pages with a formal vocabulary on the topic (ontology) and with the informal vocabulary (common terms of the topic but not in the ontology). This paper describes a web search method that takes advantage of ontologies to reduce the search area of certain topics. With this approach the relevance of search engine results is enhanced by filtering the content through the integration of domain ontologies, the WordNet thesaurus, and a hierarchical similarity measure. Thus, the improvement on the relevance of results reduces the time required to review such results.