Given is $U \sim R(0,1)$ and $X=g(U) = -\frac{1}{\lambda} \ln(U), \space \space \lambda > 0$

What's the density of random variable $X$?

$X \sim R(a,b)$ in general mean equal distribution.

Density for eual distribution is defined: $$P(a \leq X \leq b) = \int_{a}^{b} f(x) dx = \frac{1}{b-a} \int_{a}^{b} dx = 1$$

Then in this example, we have $$P(0 \leq X \leq 1) = \int_{0}^{1} g(U) dU = \int_{0}^{1} -\frac{1}{\lambda} \ln(U) \space\space dU = -\frac{1}{\lambda}\int_{0}^{1} \ln(U) \space dU = -\frac{1}{\lambda} \left[U \cdot \ln(U)-U\right]_{0}^{1}= $$

But now you see there will be problem with $\ln(0)$ when I insert intervals for calculate it. No defined calculation..

What is mistake? I don't see it :(

  • $\begingroup$ What your integral calculates is the expectation of $X$, not the density function of $X$. Your last integral can be re-expressed using the change of variables $U=\exp (-V)$; how does your concern about the $U=0$ case translate into $V$ terms? $\endgroup$ – kimchi lover Dec 6 '17 at 13:03

Use CDF transformation method. $$P(X\le x) = P(-\frac{1}{\lambda}.\ln U \le x)$$

$$= P(\ln U \ge -\lambda x)$$

$$= P(U \ge e^{-\lambda x})$$

$$ F_{X}(x) = \int_{e^{-\lambda x}}^{1} dU = 1-e^{-\lambda x}$$

$$ f(x) = \frac{d}{dx} (1-e^{-\lambda x})$$

$$f(x) = \lambda.e^{-\lambda x}$$

Hence X is $exp(\lambda)$.

  • $\begingroup$ Oh I dident know this formula thank you very much for show how do it good. Then no wonder I get this no defined result because I use wrong formula.. I have other question, what when random variable would be $\sqrt{X}$ ? Can I do it like you and in the end take square root of solution? I don't know when this make sense I only ask for interested. $\endgroup$ – roblind Dec 6 '17 at 13:16
  • $\begingroup$ You could do the same for $Y = \sqrt{X}$. with some f(x) and you want to find f(y). $\endgroup$ – Satish Ramanathan Dec 6 '17 at 13:18
  • $\begingroup$ and you are welcome $\endgroup$ – Satish Ramanathan Dec 6 '17 at 13:18
  • $\begingroup$ Ok I think I make next question for this and solve it with your idea. Thank you again :) $\endgroup$ – roblind Dec 6 '17 at 13:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.