I have an exponential distribution, and I have minitab as the company provided stat package. I am trying to have the stat package find the input values for location and shape of my particular exponential curve.

any suggestions?
without the correct exponential curve, the model won't work. . . and if i get the model to work, i get to keep my genius status at work! (genius = non math oriented person able to do a anything other than a normal distribution!)

thanks

sportsguy

The location parameter is just an offset from 0, a shortcut for adding (X) to the result. So

=ExponentialValue( 10, 1 )

would be the same as

=ExponentialValue( 0, 1 ) + 10


The shape is the hard part. Have a look at some examples on wikipedia,

[en.wikipedia.org]

If you know your data is exponential, you can use the mean. For the exponential distribution (with location 0) the mean is

1/shape

The way I would suggest approaching this would be (1) scale down the data; (2) once scaled, figure out the offset to zero, which is the location; and then use the mean of the resulting set to figure out the shape parameter.

Make sense?

Duncan Werner
dwerner@riskamp.com

Whoops, I need to correct that - we're using the scale (not the inverse scale) so the mean of the resulting distribution will be == the scale, not 1/scale.

Sorry, I know I'm just making this more confusing.

dwerner@riskamp.com