| With the development of Internet technology,the news spread more convenient.Moreover,the emergence of new media such as Weibo and WeChat has brought about the diversification of the subject and content of communication.The combination of the two has brought great changes to the news communication in the new era.Netnews is the main source of network and social public opinion.It is important to analyze and judge the influence of network news accurately.Although the current research on news influence has some achievements by scholars at home and abroad,there is no comprehensive model to measure all aspects of news influence.Most of them only consider several aspects,such as the speed of news spread,the quality of news sources,etc.Moreover,major domestic news websites don't have a clear enough model for calculating the influence index of news events.Various public opinion monitor websites only analyze the information published by netizens,and there doesn't have perfect influence monitoring system for the influence of news events.For such a situation,we need to make a visual display system from data acquisition to influence analysis based on the news event influence model.After a detailed analysis of the system architecture and the design and implementation methods of each module,confirm that the core of this research is the analysis and data mining of news event influence.Through learning the technology of web crawler,data mining,natural language processing,visual display and so on.it finally makes a visual system of news influence analysis.The main work is as follows:1.Consult the news spread literature and analyze the news influence to construct the news influence model.The definition of news influence consists of four parts:communication,credibility,persuasion and appeal,and each part puts forward a formula for calculating detailed indicators.2.A crawler program was designed for three news websites,The Chinanews online service,Tencent News and Sina News.All the news from June to December 2017 were crawled.Key information such as title,text,publishing time,news classification,author,comment number and forwarding number were analyzed and stored in MySQL database.3.Pre-processing operations such as news duplicate removal,word segmentation and removal of stop words are carried out after news data sets are acquired.Then classify the news using rules and TextCNN.4.Use Word2vec word vector and TF-IDF to determine the weight of keywords and the weight of key words of news elements to calculate the news text vector,and find out all the news data of news events according to the news text vector,then calculate each index of news influence.5.Visualize the influence of news events after calculation,and use Echarts to draw visual charts.In this research,a complete system of news influence model definition,news data acquisition,processing,index calculation and visual display is constructed. |
PDF Full Text Request