Analytical Hierarchy Method for Social Network Emergencies

Abstract

The invention provides a hierarchical analysis method for a social network emergency. The Hierarchical analysis method for the social network emergency comprises the steps that a burst hot word concurrence graph is obtained; a bipartite graph corresponding to the burst hot word concurrence graph is confirmed, a K fraction filtration processing is conducted on the bipartite graph, and all K fraction associations and each largest fraction corresponding to the correspond K fraction association are obtained; according to a preset measure index, burst hot word nodes included in each largest fraction corresponding to the correspond K fraction association are ranked in descending order; according to the rank sequence of burst hot word nodes in all largest fraction after handled after processing in the descending order, an emergency specification tree is constructed; a width precedence traversal at a K tree depth is conducted on the emergency specification tree , and K depth branches and a sub-branch corresponding to each K depth branch are confirmed; a sub-emergency of the emergency which is formed by the burst hot word nodes included in largest fraction of each branch, that is each K depth branch, and the corresponding sub-branch is confirmed, and therefore the detection for the emergency and the sub- emergency of fine-granularity is realized.

Description

technical field [0001] The invention belongs to the technical field of big data processing, and in particular relates to a hierarchical analysis method for social network emergencies. Background technique [0002] Social networks are playing an increasingly important role in people's lives, such as Weibo. The two largest microblog platforms in China, Sina and Tencent, have registered more than 500 million people. According to CNNIC's 33rd China Internet Development Survey Statistical Report, as of December 2013, the number of microblog users in my country was 281 million, and the utilization rate of microblog among netizens was 45.5%. [0003] For emergencies or hot events, Weibo's scale of influence and speed of transmission surpass ordinary blogs and traditional news media. On May 12, 2008, a major earthquake occurred in Wenchuan, Sichuan, China. Twitter disclosed the first news at about 14:35:33. Including the Linwu melon farming incident, the school bus overloading acc...

AI Technical Summary

Problems solved by technology

However, these methods generally detect based on a graph structure composed of keywords contained in each microblog data text. Since the words in the graph structure contain many redundant words for event detection, the expressiveness is not enough, which makes the microblog Poor detection results for medium incidents

Moreover, the existing event detection results list all the words corresponding to each event in the form of keyword sets, which cannot reveal the level of sub-events contained in each event, so that finer-grained event analysis cannot be achieved.

Images