Mixed Discrete-Continuous Joint Distributions
What you'll learn: How to work with joint distributions when one random variable is discrete and the other is continuous.
The Core Idea
Until now, you've likely dealt with joint distributions where both variables are discrete (using sums) or both are continuous (using double integrals). But real-world scenarios often mix types: imagine modeling the number of customer arrivals (discrete) alongside their waiting time (continuous), or the count of earthquakes (discrete) and their magnitude (continuous).
When dealing with mixed joint distributions, you combine both approaches:
- Use summation (Σ) over the discrete variable
- Use integration (∫) over the continuous variable
Notation and Structure
If X is discrete and Y is continuous, the joint distribution is described by:
- Joint PMF-PDF: p_X,Y(x, y) = P(X = x) · f_{Y|X}(y|x)
The key operations become hybrid:
Finding Marginals
- Marginal PMF of X: p_X(x) = ∫ p_X,Y(x, y) dy (integrate out Y)
- Marginal PDF of Y: f_Y(y) = Σ_x p_X,Y(x, y) (sum over X)
Finding Probabilities
To compute P(X ∈ A, Y ∈ B):
- Sum over discrete values in A
- Integrate over continuous values in B
Example: P(X = 2, 1 ≤ Y ≤ 3) = ∫₁³ p_X,Y(2, y) dy
Intuition
Think of it like organizing data in a spreadsheet where rows represent discrete categories and columns represent a continuous spectrum. To find totals, you sum across rows but integrate across the continuous dimension.
Key Takeaway: Mixed discrete-continuous joint distributions require hybrid calculations—sum over discrete variables and integrate over continuous ones, combining Σ and ∫ appropriately depending on which variable you're marginalizing or conditioning on.