Chapter

Knowledge-Based Information Retrieval and Filtering from the Web

Volume 746 of the series The Springer International Series in Engineering and Computer Science pp 285-300

Learning Ontologies for Domain-Specific Information Retrieval

  • Hele-Mai HaavAffiliated withInstitute of Cybernetics at Tallinn Technical University

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Ontologies are used in information retrieval in order to improve traditional document search methods like keyword-based search or browsing hierarchies of subject categories on the Web. To make it possible to use ontologies for that purpose requires fast automatic or semi-automatic building of formal ontologies that can be processed by a computer. This paper describes a new approach to the automatic discovery of domain-specific ontologies in order to make it possible by intelligent agents to better “understand” the intended meaning of descriptions of objects to be retrieved from different web catalogues. The approach is based on automatic construction of domain-specific ontologies using Natural Language Processing (NLP) and Formal Concept Analysis (FCA). Besides the general framework of the approach, a principal architecture of a prototypical ontology design tool OntoDesign is presented. OntoDesign is a system for automatic construction of formal domain ontologies from given domain-specific texts by using FCA.

Key words

Ontology Formal Concept Analysis Concept Lattice Learning of Concept Structures Information Retrieval