A non-parametric statistical approach for analyzing risk factor data in risk management process

Abstract

The aim of this study is to propose one practical approach to use non-parametric bootstrap technique in risk management processes especially for analyzing risk factor data, because of the fact that in most decision making cases data sizes and expert's comments are too small for analyzing risk factor data or often there are no parametric distributions on which significance can be estimated; therefore, standard statistical techniques do not always provide answers to complex risks questions. The non-parametric bootstrap is a powerful technique for assessing the accuracy of a parameter estimator in situations where conventional techniques are not valid and also non-parametric bootstrap technique is extremely valuable in situations where data sizes are too small. Bootstrap technique for decreasing the SD of risk factor data is described as well. Confidence intervals for risk factors are also obtained by means of bootstrap resampling technique. To make it more understandable, an application example is also provided. It can be concluded from the example that bootstrap will produce more accurate results in comparison with conventional techniques. © 2009 Asian Network for Scientific Information.