Automatic information extraction techniques such as named entity recognition and relation extraction have been developed but it is yet rare to apply them to various document types. In this paper, we applied them to academic literature and social media's contents in the field of diabetes to find distinctions between the perceptions of biomedical experts and the public. We analyzed and compared the experts' and the public's networks constituted by the extracted entities and relations. The results

more »

... confirmed that there are some differences in their views, i.e., biomedical entities that interest them and relations within their knowledge range.