Back to the lessonPreparing the printout…

Mathematical Foundations for Data Science · Part III — Probability and Statistics

Topic 13 · Joint Distributions and Dependence

[video] Lecture video — Topic 13: Joint Distributions and Dependence — /api/lesson-media/videos/4d3653ff-a3ca-4f63-bf8e-bb93c553e599-1794840942170-f05abd07378e7182.mp4

Flashcards · Joint Distributions and Dependence

Joint mass function — p_{X,Y}(x, y) = P(X = x, Y = y); nonnegative, and all values sum to 1.

Marginal distribution — Sum (or integrate) the joint over the other variable: p_X(x) = Σ_y p_{X,Y}(x, y).

Action rate: 0.05 + 0.15 = 0.20.

Conditional distribution — p(x | y) = p(x, y) / p_Y(y); for densities f(x | y) = f(x, y) / f_Y(y).

P(X = 1 | Y = 1) = 0.15 / 0.50 = 0.30.

Independence of variables — The joint factorizes for every pair: p(x, y) = p_X(x) p_Y(y), or f(x, y) = f_X(x) f_Y(y).

Covariance — Cov(X, Y) = E[(X − μ_X)(Y − μ_Y)] = E(XY) − E(X)E(Y); positive means they tend to move together.

0.15 − 0.20(0.50) = 0.05.

Correlation — ρ = Cov(X, Y) / (σ_X σ_Y), between −1 and 1; it measures linear association only.

0.05 / (0.4 × 0.5) = 0.25.

Uncorrelated versus independent — Independent ⇒ Cov = 0, but Cov = 0 does not imply independence.

X uniform on {−2, −1, 1, 2} and Y = |X|.

Variance of a combination — V(aX + bY) = a²V(X) + b²V(Y) + 2ab Cov(X, Y).

V(2X − Y) = 4V(X) + V(Y) − 4Cov(X, Y).

Conditional expectation — E(X | Y = y) is the mean of X under the conditional distribution; E(X | Y) is a random variable and E[E(X | Y)] = E(X).

0.8 + 0.4Y in the three-valued example.

[markdown] content available only on the platform

Mathematical Foundations for Data Science · Topic 13 · Joint Distributions and Dependence · printed on November 20, 2026