RiskResample
Description
|
RiskResample(sampMethod,{X1,X2,...,Xn}) samples from a data set with n equally likely values, the X’s. @RISK will sample from the X values using the sampling method indicated by the sampMethod index. Available sampling methods are order, random with replacement, and random without replacement.
|
Examples
|
RiskResample(2,{1,2.1,4.45,99}) returns a value from a data set with the 4 possible values 1, 2.1, 4.45 and 99. Samples will be randomly drawn with replacement from these 4 values. RiskResample(1,A1:A500) returns a value from a data set with 500 possible values. The possible values are taken from cells A1 through A500. The “order” sampling method will be used.
|
Guidelines |
sampMethod can be 1-order, 2-random with replacement, or 3- random without replacement. A RiskLibrary property function can be included with a Resample function to link the X data to a simulation output that is stored in the @RISK Library. A RiskLibrary property function causes @RISK to update the Resample X data with the current data stored for the simulation output at the start of each simulation. Therefore, if a new version of the simulation containing the output has been saved in the @RISK Library, @RISK will automatically update the RiskResample function with the new data for that output prior to simulating. When using the random with or without replacement methods, you should also consider the sampling type defined under Simulation Settings. When the number of possible values and the number of iterations are the same, all values will be used in both the random with replacement and random without replacement methods when the sampling type is set to Latin Hypercube, because each stratification will be covered exactly once. Using the sampling type set to Monte Carlo will result in duplicate values when the RiskResample sampling method is random with replacement. A RiskCorrmat or RiskCopula property function can be used with RiskResample only when the sampling method is random with replacement. RiskCorrmat and RiskCopula cannot be used with the other sampling methods. |