Stress testing incorporating causal dependency [18]

Go to: Summary | Previous | Next   
Bullet points include: If we have n Boolean random variables (E1, …, En) then number of unknown joint probabilities (e.g. each p where P(E1 = true , E2 = false, …) = p10…), is Given the joint probabilities we can always identify all marginal and conditional probabilities Given only marginal and singly conditional probabilities we cannot in general identify the joint probabilities uniquely, as the number of independent marginal and singly conditional probabilities is only: However, we can do if we assume conditional independence for the remainder

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile