首页|Study on Cross-media Hot Emergency Topic Auto-tracking
Study on Cross-media Hot Emergency Topic Auto-tracking
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
As the Internet information grows explosively, for making a scientific and efficient emergency decision in time,seeking an automatic technology to realize the effective organization and management of crossm is emergencyreports information is necessary. Thus new theoretical methods and techniques for cross-media hot emergency topic auto-tracking were proposed in this paper, effectively making the methods on simple text mode transferred into the cross-media and realizing the topic mining on both text mode and image mode information. By integrating the image visual words distribution and text words distribution into a single mode space, we madeexpression for cross-media i nformation. A combined two-step feature space dimension-reduction method was proposed in this paper, realizing the report/topic modeling in the semantic topic space. The frame and theory of the cross-media topic tracking model was given, topic tracker with information timing characteristics was constructed, and fadaptive learning of the topic model was also realized so as to track the dynamic changes in the topic. Experimental results indicated that models and methods proposed in this paper had better performance, furtherimproving the effect of hot topic autot racking.
Meiyu Liang、Junping Du、Yipeng Zhou、Yuehua Yang、Yang Song
展开 >
Beijing Key Lab of Intelligent Telecommunication Software and Multimedia,School of Computer Science,Beijing University of Posts and Telecommunications,Beijing,China
School of Computer and Information Engineering,Beijing Technology and Business University,Beijing,China
Auto-tracking Cross-media Hot Emergency automatic technology
NETL
