摘要： 设计并实现了一个Ontology 自动构建平台OntoAGS。OntoAGS 支持从给定的领域文集中抽取术语，并用领域无关文集过滤术语；支持概念学习，确定术语在其所属领域内的意义；支持实例学习，发现概念的实例；支持关系学习，找到概念间的关系。实验表明，与TextToOnto 相比，用户运用OntoAGS 能更准确有效地构建领域Ontology。

关键词: Ontology 自动构建；术语抽取；概念学习；实例学习；关系学习

Abstract: This paper presents a platform of generating domain ontology automatically, namely OntoAGS. In OntoAGS, terms are extracted from collected corpus relevant to specified domain, redundant terms are filtered by using corpus irrelevant to specified domain, concept learning methods are employed to identify the meanings of the extracted terms, and instances of learned concepts and relationship between concepts are discovered by using improved instance learning algorithm and relation learning algorithm respectively. Experiments show that OntoAGS outperforms the existing platform TextToOnto by using the above technologies

Key words: Generating ontology automatically; Terminology extraction; Concept learning; Instance learning; Relation learning