Text Visual Analysis For Web Languages

With the development of communication technology,various network platforms have emerged.As a special language form,network language has been widely used.The research on network language has many meanings,such as communication,sociology,public opinion monitoring,etc.The network language studied in this paper refers to the narrow sense of network language,that is,the communication language of Internet users.With the increasing of Internet users,the content of network language increases exponentially.The research on network language is mainly in communication and sociology,and lacks effective technical analysis of the text of network language.Based on the above background,this paper mainly studies four aspects by using text visualization technology and visual analysis method1.Large scale network language text data is obtained.Using crawler tools,the theme posts of Tianya community entertainment,stock market,economy,international observation and tourism were obtained based on the idea of opinion leaders.In order to compare the web language texts of different platforms,some news data of Sina News and Netease News were obtained,and more than 300000 web language text corpus processing and analysis were completed.2.A wordcloud generation method with strong readability is proposed.Combining with the traditional keyword selection algorithm,the paper proposes a new keyword selection method which is more suitable for the characteristics of network language by comparing word vectors.Based on the network language text,the position and co occurrence characteristics of words in the text are edited into the layout of word cloud.A word cloud layout algorithm based on double ellipse layout is proposed,which improves the readability of word cloud.3.The interactive theme River visualization method with keyword weighting is proposed.In the aspect of theme visualization,the key words are integrated into the document theme weights.The paper proposes a new method of theme River visualization which can modify the keyword weighting of the subject River fluctuation by using the weight parameters of the active words and negative words.Combined with the 3D visualization tools,a 3D interactive theme River method which can better compare the evolution of the theme trend is proposed.4.Based on the visualization method of this paper,the multi-dimensional visual analysis of network language text is completed.Combined with the above visualization methods,the paper analyzes the network language from the aspects of keywords,theme trend comparison,cross platform users' speech tendency and epidemic impact.The Internet language word cloud is used to analyze the keywords of Tianya community Entertainment Forum,and the data of five forums of Tianya community are analyzed by using interactive theme river,The paper makes a comprehensive use of various methods to complete the analysis of the tendency of cross platform network language speech and the impact of epidemic situation on network language.Through visual analysis of real text data,the analysis effect of network language text is improved,and the effectiveness of the text visualization method proposed in this paper is verified.