Research On Sentiment Analysis And Visualization System Based On Text

The rapid development of mobile Internet has promoted the development of instant messaging tools,making WeChat become a comprehensive information platform with communication,office,information,entertainment,e-commerce and so on.WeChat mainly relies on the WeChat public platform to provide information,thus giving birth to a large number of public accounts,among which major universities have also registered WeChat official accounts to convey information related to the school.Now,the analysis of tweets on WeChat official accounts generally focuses on the amount of thumb up and the number of readings.The emotional information contained in the comments is very valuable but difficult to explore.Text sentiment analysis can analyze the emotional information contained in the text,while the existing research on text sentiment analysis is mainly focused on movie reviews,Weibo,car comments,etc.,and there is a lack of research on sentiment analysis and visualization of WeChat tweet comments.This paper uses text analysis technology to mine the emotional information in tweets'comments,and uses text data visualization technology to display the analyzed information in charts.Finally,taking comment data of the school's official WeChat tweets as an example,a set of WeChat tweet comment sentiment analysis and visualization system is developed.The main research contents are divided into three parts according to the process of text data mining processing:The first is related research on data acquisition and preprocessing.The crawler algorithm is written to obtain the comment information of WeChat tweets and store the obtained data in the database.Then data preprocessing is carried out,including data cleaning,category labeling,text segmentation,removal of pause words and text vectorization.Data preprocessing not only improves the quality of text data,but also turns text data into vectors that can be recognized by computers.The data set of WeChat tweet comment sentiment analysis were made by cleaning the acquired data and marking the categories.The second is model construction.Based on literature summary,the commonly used machine learning algorithms for text sentiment analysis include support vector machine,cyclic neural network and long short term memory neural network.Divide the prepared data set into a training set and a test set,and use the training set to train these three emotion classification algorithm models.using the test set to test and evaluate the trained classification model,and comparing the classification effects of different algorithm models through experiments,The subsequent system development selects the algorithm model with better classification effect for sentiment analysis.The third is visual design.Visualization is to display the results of sentiment analysis through charts and other intuitive methods.Sentiment tendency combined with time dimension information shows the changing trend of sentiment tendency at different times.Users can interactively query the information of tweets they are interested in through the system interface,and at the same time,they can also intuitively obtain more information from the chart.This system combines text analysis technology and text data visualization technology,so that the mined information can be more clearly and intuitively displayed.Through the visualization analysis chart of this system.Through the visual analysis charts of the system,the managers of the WeChat official public platform can discover the law of tweets and emotional trends,and then play an auxiliary role in the selection of tweet topics and the supervision of public opinions.