Title :
Operational Risk Management Based on Bayesian MCMC
Author :
Zou, Qingzhong ; Li, Jinlin ; Ran, Lun
Author_Institution :
Sch. of Manage. & Econ., Beijing Inst. of Technol., Beijing, China
Abstract :
The aim of this paper is to introduce a new framework for operational risk management, based on Bayesian Markov chain Monte Carlo (MCMC). Under the LDA approach, non-conjugate distribution is used to fit the frequency and severity. One of the problems relative to the non-conjugate distribution is difficult to estimate the parameter. Then the Bayesian MCMC approach is brought forward. The Bayesian is implemented to obtain the posterior of non-conjugate distribution, the MCMC algorithm is employed to estimate the posterior parameters. The Bayesian MCMC framework is strongly recommended in the operational risk management as it incorporate internal and external loss data observations in combination with expert opinion. A numerical example is constructed to illustrate the performance of the framework advocated by this paper.
Keywords :
Bayes methods; Markov processes; Monte Carlo methods; banking; risk management; Bayesian Markov chain Monte Carlo method; LDA approach; banking; expert opinion; external loss data observations; internal loss data observations; loss distribution approach; nonconjugate distribution; operational risk management; Banking; Bayesian methods; Conference management; Frequency; Linear discriminant analysis; Monte Carlo methods; Parameter estimation; Risk management; Springs; Technology management; GB2 distribution; beyesian; loss distribution approach; markov chain monte carlo; operational risk;
Conference_Titel :
Computer Science and Information Technology - Spring Conference, 2009. IACSITSC '09. International Association of
Conference_Location :
Singapore
Print_ISBN :
978-0-7695-3653-8
DOI :
10.1109/IACSIT-SC.2009.40
