Fit Bootstrap
Parametric bootstrapping is the process by which the distribution function and it's parameters are re-sampled and refit to determine estimates for both parameter and fit statistic confidence intervals. When @RISK performs a fit with bootstrapping, the fitting process will determine the parameters for each distribution function and will then re-sample a set of data from that distribution a set number of times. These generated data sets are then refit and the results compared to the original fit to produce confidence measures of the fitted distribution's estimated parameters and statistics.
By default, bootstrapping is disabled in @RISK.
Bootstrapping can be a very time-consuming process; for each distribution being tested, that distribution will be run the number of times set in the 'Number of Resamples' configuration (see below). If testing a large number of distribution functions and bootstrapping is enabled, careful consideration should be given to the bootstrapping configuration.
Bootstrap Configuration
Figure 1 - Fit Distribution to Data - Bootstrap Tab
The Bootstrap tab of the Fit Distributions to Data window (Figure 1, right) includes the configurations for including bootstrapping in a fit.
To enable bootstrapping, check the box for 'Run Parametric Bootstrap'.
The setting for 'Number of Resamples' will determine the number of times each distribution function is run during the fitting process; please note that the higher this number, the longer the fitting process will take to complete.
The 'Parameter Confidence Level' will determine the confidence interval for the parameter values.