Applying a shift to a probability distribution during simulation | Vose Software

Applying a shift to a probability distribution during simulation

ModelRisk function: VoseShift(ShiftValue)

 

There are many possible reasons to shift a distribution. The VoseShift function can be inputted as an extra parameter into distributions (after the U-parameter) and the result is a shift of the distribution along the X-axis.

 

Example: Measurement bias

An instrument measuring voltage has a bias of -2v. The voltages being measured follow a Gamma(2,6) distribution. One can then model the observed measurements as:

VoseGamma(2,6,, VoseShift(-2))

 

The screen cap below compares the shifted and unshifted distributions:

 

Example 2: Fitting a shifted distribution to data

Many probability distributions do not support values equal to or less than zero, so cannot be fit to data with negative value unless one shifts the data by adding some constant, and then removing that shift from the fitted distribution.

 

Navigation