Fuzzy Inference for Loss Severity of Operational Risk Quantification
Guest Editor Yuantao Teng (Jilin Normal University, Shiping City, China)
To meet the Basel proposal II regulatory requirements for the Advanced Measurement Approaches in operational risk, the statistical methods of estimating operational risk technique have been explored to measure the operational risk losses in financial institutions. A fuzzy inference approach is proposed which the fuzzy parameter of lognormal distribution function is discussed in the distribution of loss severity for operational risk quantification. The distribution of loss severity for operational risk quantification can be described a lognormal distributed by fuzzy parametric statistical inference, in which parameters is characterized as a non-negative fuzzy variable. Prior membership function can be estimated using fuzzy maximum entropy rule, and then a fuzzy simulation method can be designed to estimate posterior mean. The paper shows how fuzzy variables can improve predictive performance. By simulating the operational risk loss of Chinese commercial bank and deriving the regulatory capital allotted for operational risks by China banking industry, the result shows that economical capital for each business line is in accord with the bank's asset.
Journal: International Journal of Simulation: Systems, Science and Technology, IJSSST V17
Published: no date/time given