An adaptation of Text2Onto for supporting the French language

Abstract

The ontologies are progressively imposing themselves in the field of knowledge management. While the manual construction of an ontology is by far the most reliable, this task has proved to be too tedious and expensive. To assist humans in the process of building an ontology, several tools have emerged proposing the automatic or semi-automatic construction of ontologies. In this context, Text2Onto has become one of the most recognized ontology learning tools. The performance of this tool is confirmed by several research works. However, the development of this tool is based on Princeton WordNet (PWN) for English. As a result, it is limited to the processing of textual resources written in English. In this paper, we present our approach based on JWOLF, a Java API to access the free WordNet for French that we have developed to adapt this tool for the construction of ontologies from corpus in French. To evaluate the usefulness of our approach, we assessed the performance of the improved version of Text2Onto on a simplistic corpus of French language documents. The results of this experiment have shown that the improved version of Text2Onto according to our approach is effective for the construction of an ontology from textual documents in the French language.