A Lightweight Ontology Learning Method for Chinese Government Documents

摘要

Ontology learning is a way to extract structure data from natural documents. Recently, Data-government is becoming a new trend for governments to open their data as linked data. However, there are few methods proposed to generate linked data based on Chinese government documents. To address this issue, we propose a lightweight ontology learning approach for Chinese government documents. Our method automatically extracts linked data from Chinese government documents that consist of government rules. Regular Expression is utilized to discover the semantic relationship between concepts. This is a lightweight ontology learning approach, though cheap and simple, it is proved in our experiment that it has a relative high precision value (average 85%) and a relative good recall value (average 75.7%).