Methods for detecting and summarizing emergent keywords have been extensivelystudied since social media and microblogging activities have started to play animportant role in data analysis and decision making. We present a system formonitoring emergent keywords and summarizing a document stream based on thedynamic semantic graphs of streaming documents. We introduce the notion ofdynamic eigenvector centrality for ranking emergent keywords, and present analgorithm for summarizing emergent events that is based on the minimum weightset cover. We demonstrate our system with an analysis of streaming Twitter datarelated to public security events.
展开▼