Ontology-based data mining in digital libraries

Abstract

The paper proposes matching short forms (abbreviated titles from the citation report) with their corresponding longer ones (journal titles in the digital library). The main problem is that there are often a number of syntactically different abbreviated forms for one abbreviated title in the citation report. We use character- and token-based similarity metrics to identify duplicate records. Also, we improve the process of identifying syntactically different data with the automated discovery of ontological knowledge representations such as thesauri from correctly matched data.