Research On Financial Crisis Prediction Of Listed Manufaturing Corporations With PCA-SVM Model

Manufacturing industry is at the center of industry.It is not only the engine of national industrialization,urbanization and modernization construction,but also the core subject of the national economy,or the foundation of China's technological innovation and international competitiveness.At present,China is advancing toward the “Industry 4.0” era.China's manufacturing industry is under tremendous pressure for transformation and upgrading.At the same time,China's manufacturing exports are facing international “anti-dumping” and “anti-subsidy” sanctions.The competition among domestic manufacturing companies has become increasingly fierce,leading to more and more financial crisis in manufacturing companies.Against this background,it is of great significance to establish a financial crisis warning model for listed companies in the manufacturing industry.The essence of the financial crisis is that the company continuously conducts the accumulation of financial risks and finally forms a large-scale,high-intensity outbreak.Studying the financial crisis warnings of listed companies in the manufacturing industry not only helps the company's management to monitor the company's financial status,obtains financial risks,and then takes measures to prevent financial risks,but also makes investment decisions for the company's investors,and creditors make loan decisions and The supervision of the relevant regulatory agencies can play an important role.The financial activities of listed companies in manufacturing industry will be affected by market competition,internal control,and capital management.This article selects 19 indicators based on these influencing factors and builds a financial early warning indicator system for listed companies in the manufacturing industry.This paper integrates domestic and foreign research results and constructs a financial crisis warning model based on principal component analysis and support vector machine combined model.This paper studies the samples of 27 manufacturing listed companies and similarly-sized non-ST companies that were ST in 2016 and 2017 and compares the empirical results of Logistic regression model,BP neural network model,single SVM model,and PCA-SVM model.Through analysis,it was found that PCA-SVM model has higher prediction accuracy and lower second error rate,which validates the validity and applicability of PCA-SVM model applied to the manufacturing industry's financial crisis warning.