An Analysis On Website Financial Information Of Manufacturing Listed Corporations Based On Text Mining

In recent years,the rapid development of capital market has become the main source of enterprises’absorption of funds,and many investors have poured into the capital market,hoping to obtain investment returns through the operation of the financial trading market.Thus,the business situation of the enterprise is subject to all interests and personal attention.It is especially important for investors to judge the financial situation of the company timely and accurately and avoid suffering loss of capital.Nowadays,for investors,in addition to the regular accounting data released by the company,real-time information can be viewed through the Internet to accurately understand the company’s current situation.With the rapid development of the Internet,a lot of financial information is filled with people’s life.It is very important for investors and listed companies to make use of existing technology to rapidly and accurately judge the potential value of the financial information.This paper aims at studies the large financial website published quantitative analysis of the network financial information,the network financial information index(the amount of information and the emotion value)and the correlation of the listed company’s financial situation and the profitability of listed companies,to predict the future to the stakeholders of listed companies to provide some reference information.This paper studies from two aspects:on the one hand,how to quantize the network financial information reasonably;On the other hand,the correlation between online financial information and financial situation of listed companies and the earnings status of listed companies are predicted.Around these two aspects,the specific studies the following contents:(1)the quantitative problems of network financial information.Because the network financial information is unstructured text data,therefore,it cannot directly obtain the valid information,it needs to be quantified effectively and then analyzed.The study of network financial information selects two indicators:information quantity and emotional value.For financial information network emotion value calculation,this paper adopts the method of emotional dictionary,in HowNet dictionary based on emotion,human feelings in the field of finance in new vocabulary,emotional dictionary update,make it more applicable financial sector.(2)analyze the correlation between online financial information and financial situation of listed companies.Mainly analyzes whether the network financial information related to company,mainly discusses the following aspects:financial information network is studied through correlation analysis indicators(information and emotional value)and degree of correlation,and related financial indicators.(3)Establish the profit forecast model of listed companies.In the financial index,the financial index of return on equity is selected to represent the company’s profitability,and the model of network financial information index is established..The study found that the network financial information two indicators:the amount of information and emotional value,(1).with multiple aspects of financial indicators of listed companies has relevance,normal company is both in solvency and profitability indicators and financial information is extremely significant correlation;ST(Special Treatment)is a company that has a significant correlation with financial information in terms of profitability.(2).Whether it is normal or ST company financial information network the emotional value of many financial indicators can’t express the information hidden in,in the next stage of the listed company of the company’s profitability study,emotional values can be used as a supplementary variable into the model,make the prediction more accurate.The information is not significant in this model and can be eliminated.