Beta Distributions

Sorry, the GeoGebra Applet could not be started. Please make sure that Java 1.4.2 (or later) is installed and active in your browser (Click here to install Java now) (Note: You can manipulate the shape of the curve in the above graph by dragging the handles of the "ParamA" and "ParamB" sliders.)

Beta distributions are a great way to create smooth random distributions with known limits. Beta distributions are limited to the range of 0 to 1 along the x axis, and in their normal form are also limited in such a way that the area beneath the graph is always equal to 1, or "unity".

Beta distributions are controlled by two parameters, A and B (or alpha and beta), that determine the shape of the curve. When A and B are equal and greater than 1, the distribution has a nice bell shape. When one or the other parameters is greater, then the shape is "lop-sided" toward either the left or right side. When either parameter is less than 1, the shape of the curve becomes concave instead of convex.

The first part of the formula for a beta distribution is:

	f(x) = x^(A - 1) (1 - x)^(B - 1)

You then divide f(x) by the integral between 0 and 1 of the same function to get g(x):

	g(x) = f(x) / integral(f(x), 0, 1)

Sorry, the GeoGebra Applet could not be started. Please make sure that Java 1.4.2 (or later) is installed and active in your browser (Click here to install Java now)

If, instead, you want to normalize the curve by its height instead of its area, the formula is a bit simpler. In this case you do not need to divide f(x) by its integral. Rather, you just divide it by the height of the curve at its highest point (the "vertex").

To calculate the curve's vertex, you first need to determine the x value using the following formula:

	x_vertex = (A - 1) / (A - 1 + B - 1)

Then, simply plug this x value into f(x) to get the corresponding y value. Finally, compute the following:

	g(x) = f(x) / y_vertex

To generate a random number that fits this distribution, simply plug a random value between 0 and 1 into g(x).

Download the worksheets here and here.

Listed works are licensed under a Creative Commons GNU Lesser General Public License License.

This page © Copyright 2009 Michael Horvath. Last modified: February 21 2010 20:31:53.