Commercial Bank Based on Bayesian Networks comprehensive risk early warning system
|Keywords||Commercial bank Comprehensive risk management Bayesian networks Super Bayesian methods Early Warning System|
Banking is a high-risk industry, risk prevention and management of the development of the banking sector is the eternal theme. Given commercial banks overall risk management complexity, more difficult to use traditional methods to build risk early warning system. The method is based on Bayesian network theory of probability and statistics, with a solid mathematical foundation. Because of its intuitive expression, reasoning ability, semantic clarity, and many other advantages, in uncertain reasoning has a strong advantage, so be sure theoretical research focus and in many areas has been widely used. In this paper, Bayesian networks, by constructing a comprehensive risk commercial banks topology, the various types of risk incentives of commercial banks into the overall risk of a causal association network structure, at all levels on the basis of index node assignment, the use of various types of Bayesian network indicators measure the degree of influence overall risk, and through the early warning system LED model, visually display a variety of risk factors on the overall risk of commercial banks in order to help commercial banks to take timely measures to resolve the risk, the specific research contents include: ① overview of commercial bank risk management concept of the development process, introduces the basic principles of Bayesian networks as well as its risk management in commercial banks applied research status quo. ② careful analysis of the overall risk of commercial banks, and in accordance with market risk, credit risk, operational risk and liquidity risk way to study the division of various risk sources of risk factors and metrics, and on this basis, there is under than scientifically set for each parent node child nodes to construct a complete, fully reflects the risk of commercial banks Bayesian networks. ③ The Super Bayesian method is applied to determine a priori probability Bayesian network and an example to illustrate the method a priori probability of the process. In the traditional method of expert system based on the use of mathematical tools super Bayesian approach, the individual expert opinion given different weights, a comprehensive evaluation, and ultimately get scientific priori probability values. Compared with the traditional setup method, this method does not require a huge database support and does not require complicated procedure, but it can provide a more credible Bayesian network settings prior probability, it has certain advantages. ④ completed commercial banks based on Bayesian network comprehensive risk early warning system. Meanwhile, the paper added light model, making the warning results can be reflected visually enhance the usefulness of early warning systems. ⑤ through examples, the system forward reasoning and backward reasoning can help policy-makers and risk managers to predict changes on individual factors influence the overall risk of commercial banks, as well as changes in the overall level of risk of the various factors that influence their decision-making aid and to take further measures to mitigate risk.