Looking through the application examples of a semantic web-based service, we explain how text processing can efficiently help to construct the ontology instance which is obligatorily required in establishing the service. Text processing is applied to class instantiation in the manner of identity resolution, to metadata expansion through topic and category assignment, and to object property add-up by using citation relation extraction. Consequently, 8,543 author URIs were constructed. For assigning multiple topics and categories to each literature, index terms were matched with thesaurus terms of which category names are mapped in advance. We also acquired total 135 citation networks using the citation relation automatically extracted from references in 7,237 papers. It expects that the above-mentioned methods which are put through this research will be multilaterally used in the semantic web applications.