They are called empirical because instead of taking a small set of parameters and then generating samples based on some theoretical mathematical model, they instead are free-form distributions defined by the data in the arrays. There are several “empirical” distribution function which take one or two array parameters. When the AltD version is used, any percentile values are cumulative descending percentiles. Each of the Alt forms for probability distribution functions (such as RiskNormalAlt) has a corresponding AltD version ( RiskNormalAltD). In contrast, the RiskNormal function does have a location parameter in its standard first parameter (the mean), so it does not need an extra location parameter when using an alternate parameterization.Īlternate percentile parameters can be specified in terms of cumulative descending percentiles, as well as the standard cumulative ascending percentiles. For example, the RiskGamma function does not specify a location through its standard parameter set, so a location parameter is added in its alternate parameterization. This parameter is added for distributions that do not have a built in location in their standard parameter list to allow the specification of percentiles for shifted distributions. Some distributions have an additional “location” parameter added when they are specified using alternate parameters. The standard parameter names for each distribution can be found in the heading for each function in this documentation, or in the Excel Insert Function wizard. For example, the following formula specifies a normal distribution with mean 100 and a 95th percentile of 132.89: This allows percentiles to be mixed with standard distribution arguments. If the type of parameter argument is a label in quotes (such as “mu”), the parameter is the standard distribution argument with that name. For example, the following function specifies a normal distribution with a 5th percentile of 67.10 and a 95th percentile of 132.89:Īlternate parameters can be either percentiles or standard distribution arguments (or a combination). The first specifies the type of the parameter, while the second is its value. Each parameter of an alternate parameter distribution function requires two arguments. The Alt (or AltD) versions of distribution functions provide this functionality. For example, it might be more natural to assess the 10th and 90th percentiles a normal distribution instead of the usual mean and standard deviation. refers to this as an “alternate parameter” version of the distribution. Many distribution functions can be specified with percentile values instead of their more traditional parameters. =100+RiskUniform(10,20)+(1.5*RiskNormal(A1,A2)) Alternate Distribution Functions It is possible to construct complicated formulas that combine multiple distributions. This specifies a triangular distribution with a minimum value obtained from cell B1, a most likely value 1.5 times the value in cell B2, and a maximum value obtained from cell B3. Like most Excel functions, distribution functions can have arguments that reference cells or expressions, such as Specifies that during a simulation, the cell that contains it will generate random uniformly distributed samples between 10 and 20. Probability distribution functions are used for adding uncertainty to cells and equations in a spreadsheet model.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |