Build Concept Space for Domain of Government Policy Research

Efficient search and demonstration is a hot topic in recent years. But by now, most techniques let information creeping rather than shrinking. In this paper, the government decision-making researchers´ information needs were investigated. This survey showed that the information needs of policy researchers were not the more the better, but concise. Based on the investigation, professional information sources were identified, domain specific vocabularies were established, documents were filtered and indexed. In order to improve the found probability of new concept, the co-occurrence analysis algorithms were adjusted, and the results of the analysis were carried out in a visualizing way.