Chapter 11 Probability cheatsheet
11.1 Gaussian distribution
Definition 11.1 (Multivariate Gaussian distribution) N(x;μ,Σ)=1(2π)n/2|Σ|1/2exp(−12(x−μ)′Σ−1(x−μ))
where x,μ∈Rn, Σ∈Rn×n
Linear combination Ax of Gaussian x is Gaussian with:
- mean: Aμ
- variance: AΣA′
The conditional expectation of Gaussian with respect to Ax=y is Gaussian with
- conditional mean: μ+ΣA′(AΣA′)−1(y−Aμ)
- conditional variance: Σ−ΣA′(AΣA′)−1AΣ
As a result, one obtains the following general formula:
p(Bx|Ax=y)=N(Bx;Bμ+BΣA′(AΣA′)−1(y−Aμ),BΣB′−BΣA′(AΣA′)−1AΣB′)
Note also that E(exp(λX))=exp(λμ+λ2σ22).