Research On Content And User Behavior Of Mobile New Media

With the rapid development of mobile Internet,new mobile media has popularized.Platforms such as Wechat and Toutiao generate a huge amount of content data and user behavior data.In the DT(Data technology)era,it is of great significance to analyze the characteristics of the new media platform and the characteristics of the users effectively,by using big data technology analysis and text mining technology.Firstly,this thesis summarizes the background and research significance of the new media platform.Then this thesis introduces the application of big data technology and text mining technology in user analysis.Secondly,this thesis uses HtmlUnit for data crawling and Spark for processing data.By extending the emotional dictionary and establishing sentiment score model,this thesis compares the characteristics difference between different platforms.By the method of text clustering based on ordered word vectors,we analyze audience reading preferences of new media platform.Thirdly,through hierarchical clustering based on word energy,we refine social hot topics,quantify the popularity of topics,establish topical heat cycle models,and finally obtain the spread laws of hot topics.Based on real data of new media,by the methods of text mining and big data analysis,this thesis obtains the content and user behavior research results of new media platform.Compared with the sentiment analysis of basic sentiment lexicon,the accuracy of this analysis is optimized.Compared with the text analysis algorithm of LDA,the cluster analysis method based on ordered word vectors optimizes the accuracy of article classification.Compared to other clustering methods,the hierarchical clustering method based on word energy optimizes the accuracy of hot topic extraction.These research results have important implications for the data mining of mobile new media.In addition,the research results of new media platform,such as emotional characteristics,statistical characteristics,audience characteristics,and hot topic propagation characteristics,also have great practical value.