Uji Kecocokan Data dalam Pengukuran Risiko Operasional

Abstract

Operational risk is the potential for incurring loss as a result of human error or failures in processes and control in day to day operations. On of the Basel framework provide information of AMA (Advance Measurement Approach) that can be used as quantification tools for operation risk assessment. AMA method is needed as quantification more emphasize to operational loss analysis. Goodness of Fit techniques has been developed to measure and limit deviation from probability distribution. In other words, these test show how well the distribution selected fits to the data. Distribution operational risk data can be group in frequency distribution and severitas distribution. Frequency distribution to represent discrete distribution, whereas severitas distribution represent continue distribution. Among those developed techniques that are frequently applied are Chi-square test, Kolmogorov Smirnov test and Anderson Darling test. In Goodness of Fit test, the third normality test will be compare to know which method that data fits with probability distribution. Anderson Darling test have more advantage compared with Chi-square test, Kolmogorov Smirnov test in detecting to fit data.