0
$\begingroup$

Random variable $X$ has exponential distribution with parameter $1$. What's the probability $P(X\leq2)$

How can I calculate $P(X=1)$ etc? Do I have to get aid from Poisson distribution like this:

$P(X=k)=\frac{\lambda^k}{k!}e^{-k}$

$P(X=0)=\frac{1^0}{0!}e^{-1}=\frac{1}{e}$

$P(X=1)=\frac{1^1}{1!}e^{-1}=\frac{1}{e}$

$P(X=2)=\frac{1^2}{2!}e^{-1}=\frac{1}{2e}$

$P(X=0)+P(X=1)+P(X=2)=\frac{5}{2e}$

Does it make sense?

$\endgroup$
3
  • 4
    $\begingroup$ A continuous random variable (such as exponential) has zero probability to have an exact value, so $\Pr(X=1)=0$. There are non-zero probabilities only to segments of the type $\Pr(a<X<b)$ (or $\leq$). $\endgroup$ Commented Sep 11, 2020 at 21:49
  • $\begingroup$ @YJT So I have to use intergrals like this? For example $P(X\leq 5)=\int_{0}^{5}\lambda e^{-\lambda x}dx$. How about greater? $P(X\geq2)=\int_{2}^{\infty}\lambda e^{-\lambda x}dx$? Between $P(2\leq X \leq 4)=\int_{2}^{4}\lambda e^{-\lambda x}dx$ $\endgroup$ Commented Sep 12, 2020 at 15:37
  • 1
    $\begingroup$ Yes. Or use CDF $\endgroup$ Commented Sep 12, 2020 at 16:22

1 Answer 1

3
$\begingroup$

An exponential variable is absolutely continuous and thus have a density which is a function $f$ that can be used to define the probability that the variable lies in an interval:

$$ \mathbb P(a \leq X \leq b ) = \int_a^b f(x)dx $$

For absolutely continuous variable, it is more relevant to compute the probability of the variable falling in a given interval rather than computing the probability of the variable being equal to a number because $\mathbb P(X=x)=0$ for any real number $x$.

In the case of an exponential variable with a parameter $\lambda$ we have $f(x) = \lambda e^{-\lambda x} I_{x \geq 0}$.

Thus,

\begin{align*} \mathbb P(X \leq 2 ) &= \lim_{x \to -\infty} \mathbb P(x \leq X \leq 2) \\ &= \lim_{x \to -\infty} \int_x^2 \lambda e^{-\lambda x} I_{x \geq 0} dx\\ &= \int_{- \infty}^2 \lambda e^{-\lambda x} I_{x \geq 0} dx \\ &= \int_0^2 \lambda e^{-\lambda x}dx \\ &= \lambda \left [ \frac{-1}{\lambda} e^{-\lambda x} \right]_0^2 \\ &= 1-e^{-2 \lambda} \approx 0.865 \ \text{if} \ \lambda = 1 \end{align*}

Note that since $\mathbb P(X=2) = 0$ we have $\mathbb P(X \leq 2 ) = \mathbb P(X<2)$.

$\endgroup$
2
  • 2
    $\begingroup$ Correct, but since $\mathbb P(X<0)=0$ I think it is fine to just write $$ \mathbb P(X\leqslant 2) = \int_0^2 \lambda e^{-\lambda x}\ \mathsf dx. $$ $\endgroup$ Commented Sep 11, 2020 at 23:14
  • $\begingroup$ @Math1000 So for similar tasks I only have to change values in integral? For example $P(X\leq 5)=\int_{0}^{5}\lambda e^{-\lambda x}dx$. How about greater? $P(X\geq2)=\int_{2}^{\infty}\lambda e^{-\lambda x}dx$? Between $P(2\leq X \leq 4)=\int_{2}^{4}\lambda e^{-\lambda x}dx$ $\endgroup$ Commented Sep 12, 2020 at 11:34

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.