MHB How to calculate conditional expectation E[g(x) | x>= Q] for x ~ exp(1)

AI Thread Summary
To calculate the conditional expectation E[g(x) | x >= Q] for an exponentially distributed random variable X ~ exp(1), one must integrate the function g(x) over the interval from Q to infinity, weighted by the probability density function. The function g(x) is defined as g(x) = A/exp(-bQ+c) * [(1 + exp(-bQ+c))/(1 + exp(-dx+c)) - 1]. The proposed method involves using the integral E[g[x]] = ∫_Q^∞ g(x) * f(x) dx, where f(x) = e^(-x). The solution is confirmed to be correct, utilizing the Law of the Unconscious Statistician (LOTUS) for the calculation. This approach effectively addresses the conditional expectation under the specified constraints.
user_01
Messages
5
Reaction score
0
Given that $X$ is exponentially distributed continuous random variable $X\sim \exp(1)$ and $g(x)$ is as below. How can I find the Expectectaion of $g(x)$ for the condition that $x\geq Q$, i.e. $\mathbb{E}[g(x)\ | \ x\geq Q]$.

$$g(x) = \frac{A}{\exp(-bQ+c)}\Big(\frac{1 + \exp(-bQ+c)}{1 + \exp(-dx+c)}-1\Big)$$

I suppose I should start by considering an event $\Phi$ such that $\Phi = \mathbb{P}[x \geq Q]$. However, I don't know how to go around this condition.

All constants are positive real values.
 
Physics news on Phys.org
Is the following solution correct for the above question? If it is OK, then I have found the solution. But I will really appreciate if someone can let me know if the following method is correct.

$$\mathbb{E}[g[x]] = \int_Q^\infty g(x). f(x) dx$$

$$\mathbb{E}[g[x]]= \int_Q^\infty\frac{A}{\exp(-bQ+c)}\Big(\frac{1 + \exp(-bQ+c)}{1 + \exp(-dx+c)}-1\Big) e^{-x} dx$$
 
Hi user_01,

Yes, your proposed solution is correct. The relevant theorem here is known as the Law of the Unconscious Statistician (LOTUS). See here.
 
Back
Top