The Identification And Analysis Of The Phase Event Based On The Micro-index Change Curve

With the development of Web2.0 technology, we quickly entered the era of big data. The era of big data, many news events in a variety of network platforms are reported, comments, forwarding, result in the big data of events. In the face of the big data of events, although people can obtain data about events through search engines and other tools, but these pieces of data reflects only a local event, only people from the fragmented data obtained on the basis of one-sided viewpoint effect, it is difficult to clearly grasp the process of development and change events. For this reason, the research on the identification of the events in the news topic and the evolution of the relationship between the events has become a hot topic.This thesis focuses on the discovery of phased events in news reports and the identification of evolutionary relationships between events. The following research work is carried out:(1) The hot trend of change caused by the social hot events are introduced into the identification of the events in the event. The trend of the event hot is indirectly expressed by the micro-blog’s micro-index curve, and the signal detection algorithm is used to analyze the index curve to identify the time period in which the event development process raises a wide range of attention, with the time period as a time window to organize news reports.(2)Local clustering of news reports within the time window. Based on the vector space model, the content, time and named entity feature are used to represent the news reports. The single-pass and KNN clustering algorithms are used to identify the phased events, and the clustering results are compared and analyzed.(3)Studied the evolutionary relationship between phased events.Firstly, the evolutionary relationship of the phased events is regarded as a graph structure, and then the possibility of the relationship between the two events is calculated by using the similarity of the news reports, the temporal relationship and the distribution of the reports. And the effect of the identification of the evolutionary relationship between the phased events is verified by experiments.In this thesis, we focus on the social hot event "Luo Yixiao event" to study and analyze, the experimental results show that the proposed method to identify the phased events based on the micro-index has achieved good results Finally, the evolution relation diagram of the phased events is constructed according to the identified phased events, and the evolution process of complex events is sorted out, proved the feasibility and effectiveness of this method, so as to provide a new way for the identification and evolution of events.