The Analysis And Design Of The Online Public Opinion System For Demolition And Petition Based On Machine Learning

In recent years,with the rapid development of Internet technology,it has become the main access to information.There are a lot of rumors and inflammatory statements in the midst of so many public opinions that people can't be separated from each other,which can lead some people to listen to the letter,retransmit it twice,and cause more people to be hoodwinked,so,In addition to requiring netizens to recognize the truth,network monitoring of public opinion is also very important,is the key to control the source.Especially the demolition work is related to whether the relationship between the government and the people is harmonious or not,but also to the level of happiness of the people.If it is not handled properly,it will lead to the increase of social unrest and instability factors.Therefore,this kind of network public opinion is the focus of concern,should be strictly controlled at the source,in view of improper speech timely processing.The main work of this paper is to do a network public opinion system for the public opinion related to demolition in the network.This paper first puts forward the crawling part of public opinion in this system,and then extracts specific information according to keyword crawling by compiling web crawlers for different networks,and then classifies public opinion into positive,negative and neutral emotions through the part of emotional analysis.To let the machine "understand" human language,this involves natural language processing theory,which includes pre-processing of public opinion,text cleaning,Chinese word segmentation,word vector acquisition.Finally,convolutional neural network(CNN),)is used respectively.The long short-term memory network(LSTM)and the gate recurrent unit(GRU)are used to train and predict the model.Finally,the accuracy of convolution neural network is the lowest,so that it can not be applied to the actual system.The accuracy of long-and short-term memory network is the highest,which can reach more than 85%.Finally,it is chosen as the emotional analysis model of the system.