Font Size: a A A

Research On Enterprise Financial Early Warning,Based On Random Forest-Artificial Neural Network

Posted on:2018-12-29Degree:MasterType:Thesis
Country:ChinaCandidate:Y K WangFull Text:PDF
GTID:2359330512985129Subject:Accounting
Abstract/Summary:PDF Full Text Request
Manufacturing industry has always been the leading industry of China's national economy,with the development of market economy,manufacturing enterprises are facing more and more intense competition,especially after the financial crisis.Manufacturing enterprises are facing the shortage of domestic demand,its financial risk and the possibility of a bankruptcy crisis has risen sharply,and the demand for financial early warning is increasingly urgent for manufacturing listed companies.At the same time,with the development of capital markets,other participants in the market,such as investors and regulators,are also more concerned about the financial situation of the business.Through the deep analysis of the theory of financial early warning,including financial crisis theory,financial early warning theory,macroeconomic early warning theory,we are going to find the indicators of financial early warning,so as to more scientific and comprehensive establishment of the manufacturing industry for the index system.Random forest algorithm is a kind of efficient dimensionality reduction method,which has great advantages for index screening.The neural network model can predict variables by simulating the process of biological neural network,and from non-financial perspective and Financial perspective combined to look at corporate financial early warning has a good adaptability.Combined with the random forest neural network model,the model can achieve accurate and stable early warning forecast.In the empirical process,the model T-2 year is found to reach 76%,and the T-1 year is 85%.In contrast,the artificial forest network model which based on random forest is more adaptable to the advantages of random forest and artificial neural network model.The index of the model is more scientific and the stability and accuracy of the model are improved.According to the research results above,this paper gives some prospect on the financial early warning of listed companies.Firstly,expand the scope of the study;secondly,perfect indicators-financial indicators and non-financial indicators,a comprehensive reflection of manufacturing;thirdly,comining warning-internal control and financial early warning combination can play a combination of internal and external synergies;forth,give full consideration to China's national conditions.
Keywords/Search Tags:Financial crisis, Random Forest, Artificial neural networks
PDF Full Text Request
Related items