Let {Pμ:μ∈χ} be a family of probability measures and Q a probability measure over χ. A mixture over μ is defined such that ∀A∈σ(Ω):
P(A)=∫Pμ(A)Q(dμ)
For example, we can consider {normal(0,1)μ,normal(1,1)μ} and μ∼Bernoulli(0.7). Then:
P(A)=0.7normal(0,1)(A)+0.3normal(1,1)(A)