Visualization Research On Stock Market Public Opinion Analysis

In recent years,China’s economy has developed rapidly and its international influence has been increasing.China’s stock market has become an important factor affecting global financial index.With the development of IT technology,the amount of public opinion data shows exponential growth.The traditional method of public opinion analysis can not be an effective means for investors to understand the stock market and master the operation laws of the stock market.Data visualization technology is an effective method to convey data information and further mine deep information by means of graphics.In order to facilitate the investors to understand the information,provide investors with references,and make the stock market develop healthily,this paper uses data visualization technology to visually display and analyze stock public opinion according to the characteristics of stock public opinion data.First,preprocess the stock text.Use Jieba word segmentation method to perform Chinese word segmentation on public opinion text.According to the characteristics of the stock data and the stopword list,the text data is de-stopped and the vector space model is used to represent the text.The unstructured text information is transformed into a form that can be recognized by computer.Secondly,use the XGBoost algorithm to classify stock public opinion.In order to further improve the classification accuracy,an improved XGBoost algorithm is proposed,which uses the Bayesian optimization algorithm to make full use of prior knowledge to continuously approximate the function distribution.Adjust the hyperparameters and gradually find the parameters that maximize the global improvement,and bring the optimal hyperparameters into the XGBoost algorithm for text classification.Again,a bubble treemap is introduced for the hierarchical structure of stocks,and an interactive visualization design based on a bubble treemap is proposed.The bubble treemap reflects the hierarchical characteristics of stock data.The interactive visualization design realizes the visualization of stock information and the enlargement of different levels of nodes,and displays the stock public opinion information from multiple angles.Finally,according to the above method,the data of Shanghai and Shenzhen 300 constituent stocks were tested,and the experimental results were visualized,which verified the validity of the method in this paper.