Calculate Alpha, Beta for use in Beta Dist function

Hello

I wish to calculate alpha and beta for use in the Beta Dist function for a non normal distribution.

Having searched the web and various fora I can only find formula used for calculating alpha and beta for normally distributed values.

eg

http://www.xycoon.com/beta.htm

I am looking for a way of calculating/estimating alpha and beta for skewed distributions. The explanation in Excel seems to assume that you already know what alpha and beta are.

Do you know of a way of calculating them in these circumstances?

Regards

Andrew

I have used to Solver and the =BETADIST() function to estimate alpha and beta based on raw data. In essence, it was basically done like a non-linear regression:

1) Guess a value for alpha and beta.
2) Use the function to calculate the probability or cumulative density at the various known_x's
3) Compare the calculated values to the "measured" values. I used a basic least squares type objective function [minimize sum((yestimated-ymeasured)^2)], though you could use whatever objective function seems most appropriate to your situation.
4) Call Solver and tell it to minimize the sum of squares (or whatever you've chosen as your objective function) by changing alpha and beta.
5) Evaluate the "regression" to see if the returned alpha and beta are reasonable.

Many thanks for your response.

I haven't used Solver before so I'll have to look at that tool first.