Font Size: a A A

Statistical Analysis Of Financial Distress Warning Of China’s Manufacturing Listed Enterprises

Posted on:2017-05-28Degree:MasterType:Thesis
Country:ChinaCandidate:J X HanFull Text:PDF
GTID:2309330482488575Subject:Statistics
Abstract/Summary:PDF Full Text Request
With the competition increasing severely, more and more China’s manufacturing listed enterprises fall into trouble. However, financial crisis is one of the most important reasons. Therefore, it is very essential to build a financial crises warning system for China’s manufacturing listed companies. The building of financial crises warning system is not only significant for the management of the companies, but also for the investment decision making of investors, creditors, banks and supervisory of financial authority.China’s manufacturing listed enterprises in 2014 are divided into 3 categories based on its current financial conditions.35 companies with *ST in 2014 is financial-crisis-condition companies (Y=-1). The companies without *ST is divided into 2 groups by Altaian’Z-score models. The company which score is more than 1.8 is defined as well-financial-condition company (Y= 1), and the company which score is less than 1.8 is defined as grey-financial-conditions company (Y= 0). Each 35 random companies are extracted from all the well-financial-condition companies and grey-financial-condition companies. Then the 18 indexes in 2012 and 2013 are chosen from RESSET. Through Kruskal-Wallis H test, some indexes are rejected, and the left indexes are put into factor analysis. Each 25 random companies are chosen from 35 companies as training samples, other 10 companies are test samples. The training samples and test samples are both used in discriminant analysis, ordered logical regression analysis, SVM model and combined BP neural network model to forecast the financial conditions in 2014. From the aspect of data, the result shows the data of 2013 is better. From the aspect of accuracy and simplicity, SVM model is much better. However, the result of combined BP neural network model is superior when the amount of data is larger. The investors, creditors, banks and financial authority can use the models studied in this paper to supervise the financial conditions of China’s manufacturing listed enterprises.
Keywords/Search Tags:Financial crisis early warning, Discriminant analysis, Ordered logical regression analysis, SVM, Combined BP neural network model
PDF Full Text Request
Related items