Burst Event Detection And Initialyzing Technology Research In Micro-blog

With the rapid development of the network, the network gradually integrated into people's lives.Microblogging is also developing rapidly in recent three years, Emergency monitoring and traceability has become a hot research direction. Burst event detection technology and event initial detect technology help people to detect incidents automatically.It helps people learn more about the causes and consequences of emergencies. Firstly we studied about chinese segmentation,message spread model,event spread model,Burst detection model and the user relational model. then,we give a event detection algorithm based on MACD emergencies and a multidimensional analysis method based on microblogging emergencies traceable.This paper firstly presents a new segmentation method which is a combination of dictionary-segmentation method and Statistical-segmentation method.It is Effective and accurate.This paper proposes an event detection algorithm based on MACD, consider using user influence and user event schedule sliding window mechanism, with which, improve the accuracy of event detection algorithms from two aspects. When clustering event, this paper use a algorithm based on burst-word distance. It's proved efficient. Experimental results show that, the algorithm can effectively remove users have rest and low-quality users to improve the accuracy of detection of emergencies.This paper presents microblogging emergencies traceable method based on multi-dimensional analysis. The method works on the microblogging property dimensions,structural dimensions to calculate initial weight of each microblogging. Then we calculate microblogging final weights based on analyzing the relationship between the behavior-based microblogging dimensions. This paper also presents similarity calculation method which based on the contents of the burst words, compared with the text edit distance calculation method,it runs faster.This paper proposed analysis methods to measure the weight between users effectively after studies the interaction between users. Experiments show that the method has a lower time complexity, higher traceable accuracy.