Knowledge Discovery from Text Learning for Ontology Modeling

This paper presents a methodology of knowledge discovery from text learning for ontology modeling. Knowledge written in text is always hard to be extracted by automated process, and most existing ontologies are defined manually. Those ontologies are not comprehensive enough to express most human knowledge in the real world. Therefore, the most efficient way to identify knowledge is discovering it from rich text. In this paper, we proposed a statistical based method to measure the relation of appearing frequency of word in text. The method identifies and discovers knowledge by automated process. We also defined ontology model - ontology graph, to express knowledge, the graph facilitates machine and human processing. The extracted knowledge in the graph format can aid user to revise and define ontology knowledge more effectively and accurately.