Bayes Estimation Of Operation Risk Measurement Model In Securities Industry

Mathematical statistics is to find out inner regularity of some phenomena by observing frequency of these phenomena and perform judgment and prediction with certain accuracy. Carry out induction and reorganization to some results of these researches and certain mathematical schemes are formed gradually. Internationally, mathematical statistics are usually divided into two categories: classical school and Bayes school. Their difference is whether they use apriori information. Concept of bayes estimation was firstly put forward by Robbins in 1950 s. After development of more than a century, bayes method has gradually developed into a very important school in statistics. Bayes method is to estimate some features of unknown parameter according to existing data. This process mode makes result better and closer to life, which indicates applied universality of bayes estimation.Operational risk plays an important role in security industry due to instability of frequency and human factor, which makes it difficult to predict operational risk through quantization. How to quantify operational risk and add quantized operational risk to operational risk management system in security industry has become one of the important contents researched by researchers in recent years. Primary cause of difficulty of quantization of operational risk is data deficiency and it is difficult to be collected. Lack of loss data will directly affects estimation of parameter, thus making estimation inaccurate.One of the bright points in this paper is to confirm parameter of model through bayes method, thus effectively solving a series of problems caused by data deficiency.Main contents of this paper: the first chapter briefly introduces bayes estimation and based content of operational risk; the second chapter briefly depicts decision-making theory and nature of bayes estimation; the third chapter confirms parameter of loss data by constructing model and making use of estimation of maximum likelihood and bayes estimation, and gains the result of loss data of operational risk and makes comparison. The fourth chapter finds Bayes priori estimation of change point of independent random variables sequence under asymmetric loss functions in process of performing bayes estimation to parameter, and carries out analytical investigation to it.