Instance Learning in Automatic Construction of Domain Ontology
Ontology自动创建中实例学习的研究

The importance of Ontology is being recognized in various research and application areas. Constructing domain Ontology takes too much time and manpower. Hence, it is of great significance to do some research on constructing domain Ontology automatically. Instance is an important element of Ontology. Learning instance from domain corpus is a key issue in automatic construction of domain Ontology. A domain Ontology construction system OntoAGS has been developed. Given some domain corpus, OntoAGS can construct domain Ontology automatically. The instance learning method of this system is based on pattern matching algorithm. Experiments show that OntoAGS can generally learn more instances and more accurately than text-to-onto system which is a popular Ontology semi-automatic construction system.