The research of concept extraction in ontology extension based on extended association rules

Abstract

Ontology extension is to achieve the purpose of expanding ontology by adding new concepts and the relationship between concepts to the existing ontology. In this paper, we proposed a new method to automatically extract concept from text collection and attach them with existing ontology. We mainly use the weighted association rules to mine latent concepts from the unstructured text. Then, according to the TF-IDF value of the concept, we determine a set of the weights of the weighted association rules. At the same time, on the basis of the semantic similarity between concepts in the initial ontology to be expanded, the minimum support in the association rules is determined. This method can improve the efficiency and accuracy of domain concept extraction. We compare our new method with TF-IDF on data sets. Experiment results show that our method outperforms TF-IDF in terms of accuracy and provides better ontology description.