💡 Today I Learned

Weibull function

This is a handy function for modeling statistical likelihood of failure (or survival).

There are four parameters:

WEIBULL(x,alpha,beta,Cumulative) gives a number between 0 and 1.

X is the value at which to evaluate the function.
Alpha is a parameter to the distribution.
Beta is a parameter to the distribution.
Cumulative (true or false) determines the form of the function.

Here is Weibull function for the cumulative and non-cumulative case.

WEIBULL(x,a,b,TRUE) = 1 - e^(-(x/b)^a)
WEIBULL(x,a,b,FALSE) = (a/b^a)(x^(a-1))(e^(-x/b)^a)

In Excel, there is a function WEIBULL, which you can use directly.

If you need to re-implement it in SQL Server, here it is for the cumulative case...

ALTER function dbo.WeibullCumulative(@x float, @alpha float, @beta float)
returns float
as
BEGIN
 return 1 - exp(-1 * POWER((@x / @beta), @alpha))
END

And I haven't bothered with the non-cumulative, but it is similar, I think...

(@alpha/Power(@beta,@alpha))* Power(@x, @alpha-1) * exp(POWER((-1 * @x / @beta), @alpha))

Though I haven't tested it. So be careful with that and tell me if you do implement it.

Source