Probability at least one, using correlation

1k Views Asked by At

I have a problem using the correlation in combination with the "at least one" probability.

I have $P(A)=57\%$, and $P(B)=74\%$, and I calculated their correlation coefficient and it is $0.1557$.

To calculate the $P(\text{at least one})$ I did it this way:

$$P(\text{at least one happens}) = 1 - P(\text{neither A nor B happens})$$

for which I did $P(A)' = 43\%$, $P(B)' = 26\%$,

$$P(\text{neither A nor B}) = P(A)'P(B)' = (0.43)(0.26) = 0.1118.$$

So using the first formula: $$P(\text{at least one happens}) = 1 - 0.1118 = 0.8882.$$

So the probability of at least one of them happening is $88.82\%$, however, I need to combine this with their correlation and I'm having troubles doing it that way, I would appreciate it if someone could help me with this problem.

Thank you!

PD. Forgot to add their are dependent events.

1

There are 1 best solutions below

2
On BEST ANSWER

What does the correlation coefficient of two events mean?   Is it the correlation coefficient of their Bernoulli indicator random variables?   Then

$$\newcommand{\Chi}{{\raise{0.5ex}{\chi}}} \begin{align} \rho_{\!\lower{0.5ex}{A,B}} & = {\sf Corr}(\Chi_A, \Chi_B) \\[1ex] & = \dfrac{{\sf Cov}(\Chi_A, \Chi_B)}{\sqrt{\,{\sf Var}(\Chi_A){\sf Var}(\Chi_B)\,}} \\[1ex] & = \dfrac{{\sf E}(\Chi_A\Chi_B)-{\sf E}(\Chi_A){\sf E}(\Chi_B)}{\sqrt{\,({\sf E}({\Chi}_A^2)-{\sf E}({\Chi}_A)^2)\,({\sf E}({\Chi}_B^2)-{\sf E}({\Chi}_B)^2)\;}} \\[1ex] & = \dfrac{{\sf P}(A\cap B)-{\sf P}(A){\sf P}(B)}{\sqrt{\,({\sf P}(A)-{\sf P}(A)^2)\,({\sf P}(B)-{\sf P}(B)^2)\;}} \\[2ex] \therefore {\sf P}(A\cap B) & = {\sf P}(A){\sf P}(B)+ \rho_{\!\lower{0.5ex}{A,B}} \sqrt{\,{\sf P}(A)(1-{\sf P}(A))\,{\sf P}(B)(1-{\sf P}(B))\;} \end{align} $$