Warning Model of Financial Risk Literature Review

Abstract: Based on the domestic Financial Risk early warning model for the introduction of the existing Research results and literature were summarized, and finally with the progress of domestic Research on early warning of Financial Risks to make the evaluation model.

Keywords: Domestic Financial Risk Model
0 Introduction
Financial Risk early warning is provided through corporate financial statements, business plans and other relevant accounting information, the use of accounting, statistics, finance, business management, marketing theory, the use of ratio analysis, comparative analysis, factor analysis and multiple analysis methods, business activities, financial activities for analysis and forecasting, to find the management activities of enterprises in the potential Financial Risks, and before the crisis issue warning to business, business management urged the authorities to take effective measures to avoid the potential risk change way into the loss of a warning.

Warning of the domestic financial crisis began in the late 20th century, 80, on the financial crisis early warning model is up to the late 20th century, began to 90. Wushi Nong, Huang Shizhong (1986) has written <"China Economic issues>> Introduction Enterprise bankruptcy of the financial analysis and forecasting models: the National Natural Science Foundation of Management Science Group, who has support in corporate SHE Lian early warning Research, and published in 1999 warning of enterprise management books, scholars in China began in earnest after the financial early warning of , and have achieved certain results.

A statistical prediction model
1.1 Single variable prediction model. Single variable prediction model refers to a financial index as a criterion to judge whether a company is in bankruptcy prediction model state.

Chen Jing (1999 to 27 ST companies and 27 non-ST companies as a sample, the final selection of asset-liability ratio, return on net assets and other financial indicators 6, ST respectively the year before the company was two years ago, the former three years of financial data, discriminant analysis made use of empirical Research. In univariate analysis, found in the debt ratio, current ratio, return on assets, return on net assets four indicators, the current ratio and debt ratio the lowest false positive rate.

More than 1.2 variable prediction model. Multivariable models using various financial ratios that summarize the structure weighted multiple linear function formula to predict the financial crisis.

The first week of China, Yang Jihua and Wang Ping (1996) Z scores improved on the basis of the model, taking into account changes in cash flow targets, establish a model of F scores:
F =- 0.1774 +1.1091 X1 +0.1704 X2 + l.9271X3 +0.0302 X4 +0.4961 X5
Where, X1, X2 and X4 and Z score model in the X1, X2 and X4 indicators reflect the same, while the X3, X5 and Z Score model X3, X5 different. X3 = (after-tax net income + depreciation) / average total liabilities, it is a cash flow variable is a measure of corporate cash flows generated by all the debt can be used to repay an important indicator of business. X5 = (after-tax net income + interest + depreciation) / average total assets, the determination of the enterprise Total assets in the ability to create cash flow (where the interest is business interest income to interest payments the balance). relative to the Z score model, it can more accurately predict the existence of financial risk business.

Zhang Ling (2000) to 120 companies for the study, 60 of which use the company's financial data to estimate binary linear discriminant model, and use the other 60 companies to test the model, that model has four years ahead of predictions . In addition, the use of academics as well Huangyan MDA modeling and Liyuan Xu (2001, Yin Xia et al (2001, to Dewei (2002, Wei Jianguo, etc. (2002, Yang Shue and Xu Weigang (2003, Tang Zhenyu et al (2004 and Ben Friends of Red ( 2005.

1.3 Logit.Logit model is a series of financial indicators used to predict the probability of financial crisis, then the bank, investors risk preference settings such as the risk warning line, as the object of analysis and decision making for risk positioning.

Wu Shinong, Luxian Yi (2001 selected 70 companies in financial trouble and 70 samples of normal financial company, the first Application of profile analysis and univariate analysis to determine financial distress before the onset of the two types of company within 5 years of 21 per financial year index difference, the final selection of six financial indicators as early warning indicators, applied Fisher linear decision analysis, multiple linear regression analysis and Logistic regression analysis are three types of early warning models were established, and that Application of Logistic Regression Analysis Method the lowest false positive rate of early warning models. Links http://www.hi138.com Research Papers Download
In addition, Jiang Xiuhua and Sun Zheng (2001 to discuss the best split point that the probability of 0.1 as the best split point, Qiaozhuo (2002 and Qi Zhiping (2002 introduction of quadratic terms and cross-term model, Chen and Chen Zhihong (2000, Song force and Ulrich (2004 adjusted financial data modeling, Zhang Ming and Cheng Tao (2005, Liang Qi (2005, Zhang Yang (2005 through the use of principal component analysis to reduce the dimensions of the logistic method, a solution of linear problems was modeled, Gu silver wide (2005 A model based on Jackknife test are predicted to improve the accuracy of the model.

1.4 Kernel methods. Luoyou Xi et al (2005 through the reduction of the principal component analysis model indicators, using kernel function modeling, the results show that the model performance than traditional forecasting methods, can solve the problem of large-scale Application of the sample set.

Intelligent Early Warning Model 2
Intelligent early-warning model is mainly based on neural network analysis of the various models. Neural network model is an artificial intelligence system to simulate biological neural system model, the use of the training process repeated so that the accumulation of their own to achieve learning through experience effect.

INDUSTRIAL Wan Hai Hui, Zhang Wei et al (1999) used neural networks of commercial banks, financial risk and found that neural network has strong nonlinear mapping ability, the ability of their learning experience: academics pair analysis, etc. (2001) analysis of the BP neural network method applied to the bank's financial early warning analysis, built financial early warning of non-linear model.

Liu, He Guangjun (2004 to 728 samples, 36 financial indicators for early warning of financial crisis. They are the traditional discriminant analysis and logistic regression analysis, based on artificial neural network to explore the financial crisis early warning. Results showed that the artificial neural network prediction accuracy rate is higher than the first two methods of forecasting accuracy rate. Xieji Gang (2004 classification and others use an integrated approach to financial distress, the results found that the prediction accuracy rate can reach 86%.

Zhang Genming to Xiao Ji and Sun Jingyi (2006 with BP neural network method, the 263 listed companies in the manufacturing sector as a cross-sectional study sample of financial indicators, and using 76 as the test sample of listed manufacturing companies, listed companies in the manufacturing sector to establish financial crisis early warning. The results show that there is no distinction between the industry and the general financial distress models, the sub-sector financial distress BP neural network model prediction accuracy rate was higher for the majority of investors and financial forecasting firm regulators provide a more reliable basis.

In addition, the Pre Sun Yi (2006 proposed based on the rough - fuzzy neural network (Rough-Fuzzy-ANN model and the corresponding algorithm, through the financial data of listed companies in China based on the empirical analysis shows that This model has high accuracy, learning and generalization ability, and adaptability advantages for enterprises to provide early warning of financial crisis, a new dynamic way. Zhang Lin (2004 CBR technology to build a business using financial early warning system for degree of enterprise financial crisis, police monitoring, warning sign identification and anti-police and other police row provides a new idea.

3 and Comparative Study of Mixed Mode
In recent years, there was even a financial early warning of mixed mode. Hybrid model is also using two or more kinds of methods to model for financial early warning analysis. This were empirical studies have shown that the hybrid model method model than a single-phase type has a higher accuracy. such as: Xu Yong (2007 listed companies in the manufacturing sector in Shanghai and Shenzhen as a sample, determined by Fisher linear analysis and Logistic regression analysis of listed companies in China manufacturing model of financial distress. The study took into account differences in industry-financial indicators, selected as a sample of listed manufacturing companies, will enable more targeted results.

4 Early Warning Model of Financial Risk Assessment
Through the domestic financial distress model, from a single variable, multivariable, Logit model based on these statistical methods to study the neural network model of intelligence and other non-statistical model, from a single model to the hybrid model and its comparative study from the financial indicators-based Research to study the introduction of non-financial indicators, financial distress model by domestic practitioners and academic attention and has made significant progress. reposted elsewhere in the Research Papers Download http:/ / www.hi138.com
At the same time, the domestic research in this area there are still shortcomings. ① early warning model for different conditions are different, the model will inevitably affect the accuracy and precision. The current financial crisis early warning model for most of the emphasis on using financial indicators to build models, rather than indicators of financial factors, not only in data collection difficult and easy evaluation. ② domestic scholars study these models are obtained through empirical studies, lack of theoretical guidance, the researchers in the choice of variables, they also by their own the impact of value judgments. ③ existing financial early warning in the country studies, the selection of indicators are often based on general financial theory, risk theory and management theory, and sometimes depends on the researcher's intuitive judgments, and the availability of data, many studies who have not found a convincing theory of financial risks and to support the establishment of early warning early warning model, which resulted in a variety of financial early warning early warning model is inconsistent conclusions.

Our research in this area and scale of the industry on how to consider the impact of changes in conditions of sample design can affect the prediction accuracy, theoretically and empirically how the introduction of more effective predictor variables, and indicators of domestic research on the importance attached cash flow is not enough These are our future direction for further research.

References:
[1] Jiang Xiuhua, Ren Qiang, Sun Zheng. Listed companies financial distress model [J]. Forecast .2002.3.

[2] Zhang Ming. Forefront of corporate financial early warning research. Beijing. China Financial and Economic Publishing set .2004.

[3] Zhao Ying. Forecasting Model and Its Application in China reviewed. HKSAR Economic .2005
[4] cool. Corporate financial risk early warning mechanism. Modern commercial .2009
[5] Zhang Wei. Enterprise Financial Early Warning Model in the use of cash flow analysis. Modern Finance .2007. Links http://www.hi138.com Research Papers Download