Design And Implementation Of Microblog Burst Topic Detection System Based On Propagation Value

With increasingly development of information technology,web era has slowly come into the life of Internet users,since the media platform is popular with everyone,and micro-blog is a typical example.The rise of the media brings new experiences to people,users can publish photos and text messages and even video and other multimedia information,and share the latest trends of the individual,the side of the real-time information to share with others by some ways of communication,such as mobile phone,QQ,We-Chat,email,web and so on.It is for this reason,in micro-blog,users have new features,manufacturing information and dissemination of information.The distribution and forwarding of the information is convenient and fast,so micro-blog will occupy the leading position in information propagation speed.On the micro-blog platform,the sudden appear on the social subject propagation way is to use the attention of the relationship between micro-blog users to forward.This way of communication is easy to extend to the traditional media with a huge impact on society,the most common traditional media is news.As a result,the emergent topic detection technology based on micro-blog is positive on practical aspects,such as discovering the social hot spots,perception of network public opinion as soon as possible,reflecting to emergencies as soon as possible,etc.Based on this,this paper will design and implement a system of micro-blog burst topic detection based on promulgating value,mainly focuses on the following three aspects of the study:?(1)In the preparation stage of this paper,I have read a lot of documents and relevant materials about the research situation of topic detection so far,on the basis of understanding the basis of burst topic detection,demand analysis is identified in this paper.And in this system,there are 3 kinds of function to be achieved: information collection,data processing,and burst topic detection.?(2)Based on the basis of the whole process of the analysis of topic detection,considering of the characteristics of micro-blog,such as short text,real-time,social networking,media,and so on,to ensure that the two characteristics of sudden word also reflect the degree of differentiation of abrupt words,so promulgating value is defined and introduced.The process of its calculation includes two aspects: the influence of users and relative superiority.An algorithm of emergent topic detection based on promulgating value is designed,the first step,the candidate burst words are determined according to the changes of time,the second step is using weighting and promulgating value to calculate complete sudden word filter.Lastly,word occurrenceprobability is used to calculate a sudden similarity between words,using the single-pass clustering algorithm to complete topic detection.(3)It is designed a work flow of the emergency of topic detection system based on dissemination on the basis of system requirements analysis.On the whole system architecture,analyzing of two models respectively C/S and B/S,B/S development mode is more suitable for the design of this system;In the design of the system level,the system is in compliance with the principles of high cohesion and low coupling,the system is using three-tier architecture: the data access layer(DAL),business logic layer(BLL),the presentation layer(UI).In addition,the system's choice of database is Microsoft SQL Server 2008 version,development language is c #.