How to make Ontologies self-building from Wiki-Texts

While it used to be satisfactory that sea r ch applications used to look up simple keywords in the past, nowadays a sea rch engine is increasingly expected to be able performing sema ntic sea rch as well. This leads to an intensified research on ontologies supporting these applications. Users might only have a vague idea of the exact keywords to the content they seek, especially when it comes to technical or subject-specific topics and professional expertise is needed. A domain-specific ontology can support the user suggesting differentiated or alternative keywords for his query. Nevertheless, building an ontology containing expert knowledge is effortful and domain experts normally do not have sufficient skills how to build an ontology. Therefore, it is desirable to make an ontology automatically build and extend itself from alrea dy ava ilable information, such as wiki-texts. We developed a text mining component which automatically derives an ontology from collaborative texts written by scientific experts. Our component is used in a sea rch application of an informa tion system on energy research.