Course contentsShow
← Mathematics
Lesson 2870 of 3,12156. Probability — Random VariablesFree lesson

The Jacobian Method for Transformations

Apply the change-of-variables technique with Jacobian to find PDFs of transformed variables.

The Jacobian Method for Transformations

What You'll Learn

You'll learn how to find the probability density function of a transformed continuous random variable using the Jacobian determinant when changing variables.

The Core Idea

When you transform a continuous random variable through a function, its PDF doesn't simply "plug in" the new variable. The transformation stretches or compresses the probability density, and the Jacobian measures exactly how much this distortion occurs.

Think of it like converting currencies with an exchange rate, but the "rate" changes depending on where you are in the distribution.

The Method

Suppose you have a continuous random variable X with PDF f_X(x), and you transform it: Y = g(X).

To find the PDF of Y:

  1. Invert the transformation: Solve for x = g⁻¹(y)
  2. Compute the Jacobian: Find the derivative dx/dy (how x changes with respect to y)
  3. Apply the formula: f_Y(y) = f_X(g⁻¹(y)) · |dx/dy|

The absolute value of dx/dy is called the Jacobian, and it accounts for how the transformation "rescales" probability density.

Why the Jacobian?

Probability must be conserved. If a transformation compresses an interval (making it shorter), the density must increase proportionally so the total probability stays the same. The Jacobian captures this stretching factor.

For multidimensional transformations (say, X₁, X₂ → Y₁, Y₂), the Jacobian becomes a determinant of partial derivatives, measuring how area (or volume) changes under transformation.

Example Intuition

If Y = 2X, then an interval of length 1 in Y-space corresponds to length 0.5 in X-space. The Jacobian |dx/dy| = 1/2 accounts for this compression, ensuring probability is properly redistributed.

Key Takeaway: The Jacobian method uses the derivative (or determinant) of the inverse transformation to correctly adjust probability density when variables are transformed, ensuring total probability remains 1.