QA4 - Multivariate Random Variables Flashcards
Compute the marginal and conditional distributions of a discrete bivariate random variable
Marginal is sum across the row / column
Conditional is divide prob of realisation / prob of overall realisation
Define covariance and explain what it measures
Is a measure of dispersion and describes how variables move together
Explain the relationship between covariance and correlation of two random variables and how these are related to the independence of the two variables
Two variables must have zero correlation but not necessarily independent if zero correlation
Covariance of two independent variables is also zero
Explain the effects of a linear transformation on the covariance and correlation between two random variables
Covariance scales by multipliers
Correlation changes by ab / (mod(a) * mod(b)) where a and b are multipliers
Compute the variance of a weighted sum of two random variables
a^2 * Var(X1) + b^2 * Var(X) + 2ab * Cov(X1, X2)