Objectives Earthquake occurred frequently in China. More knowledge needs to be discovered to take timely emergency actions to mitigate earthquake damage. Thus the construction method of knowledge discovery model for earthquake emergency is one of the core research issues in earthquake emergency field. It is worth studying how to realize knowledge discovery for earthquake emergency from abundant and complicated data under less support of prior knowledge.
Methods We put forward a kind of text big data based knowledge discovery model for earthquake emergency. First, text data associated with earthquake emergency is collected, including both academic document data (CNKI(China national knowledge infrastructure) data) sets and social media data sets(Weibo data). Then CiteSpace tool and formal concept analysis method are utilized to extract the high-frequency keywords and their linkages, the linkages’ frequency between keywords as linkage strength, a complex network of earthquake emergency knowledge is established for community classification. After community dividing, several big communities can be divided from the former established complex network.
Results From big communities’ further analysis : (1) in seismic resistance design domain, the discovered knowledge has high coherence with expert knowledge; (2) in earthquake rescue domain, the discovered knowledge shows that China Earthquake Administration have responsibility to afford emergency information services as well as public opinion guiding; (3)in geological hazards domain, the discovered knowledge founds the vegetation restoration in geological disaster area should be paid more attention; (4) the knowledge discovered from academic document data and social media data can be complementary.
Conclusions The experiment result shows that the model can find relevant knowledge, especially those experts rarely concern.