Below is a problem I did. However, I did not come up with the answer in book. I am thinking that I might have the wrong limits for the integral. I am hoping somebody can point out what I did wrong.



Let $Y = \sin X$, where $X$ is uniformly distributed over $(0, 2 \pi)$. Find the pdf of $Y$.


\begin{eqnarray*} P(Y<=y_0) &=& P( \sin x <= y_0 ) = P( x <= \arcsin( y_0) ) \\ P(Y<=y_0) &=& \int_0^{\arcsin(y_0)} \frac{1}{2 \pi} dx = \frac{x}{2 \pi} \Big|_0^{\arcsin(y_0)} = \frac{\arcsin(y_0)}{2 \pi} - \frac{0}{2 \pi} \\ P(Y<=y_0) &=& \frac{\arcsin(y_0)}{2 \pi} \\ F(Y) &=& \frac{\arcsin(Y)}{2 \pi} \\ f(y) &=& \frac{1}{2 \pi \sqrt{1 - y^2}} \\ \end{eqnarray*} However, the book gets: \begin{eqnarray*} f(y) &=& \frac{1}{\pi \sqrt{1 - y^2}} \\ \end{eqnarray*} I believe that book is right because \begin{eqnarray*} \int_{-1}^{1} \frac{1}{\pi \sqrt{1 - y^2} } dy &=& \frac{\arcsin{y}}{\pi} \Big|_{-1}^{1} = \frac{\frac{\pi}{2}}{\pi} - \frac{-\frac{\pi}{2}}{\pi} \\ \int_{-1}^{1} \frac{1}{\pi \sqrt{1 - y^2} } dy &=& 1 \\ \end{eqnarray*}

  • $\begingroup$ Note that $\arcsin$ is defined only over $\left[-\frac{\pi}{2}, \frac{\pi}{2}\right]$, so writing $P(\sin(x) \leq y_0) = P(X \leq \arcsin(y_0))$ is invalid, since $y_0 \in [-1, 1]$. $\endgroup$ – Clarinetist Mar 25 '16 at 13:27
  • $\begingroup$ I am still confused. If I had written $sin ^{-1}(y)$ would that have been right? Do I need to write several integrals? $\endgroup$ – Bob Mar 25 '16 at 13:52

If $X$ is uniform in $(0,2\pi)$, then has cdf $$ F_X(x)=\begin{cases} 0 & x\le 0\\ \frac{x}{2\pi} & x\in (0,2\pi)\\ 1 & x\ge 2\pi \end{cases} $$ The random variable $Y=\sin (X)$ takes values on $(-1,1)$. Hence, $\Bbb P(Y\le y) = 0$ for $y \le -1$ and $\Bbb P(Y\le y) = 1$ for $y \ge 1$. Let now $y \in (-1, 1)$. We have $$ F_Y(y)=\Bbb P(Y\le y)=\Bbb P(\sin (X)\le y) $$ The equation $\sin(x) = y$ has two solutions in the interval $(0, 2\pi)$:

  • $x = \arcsin(y)$ and $\pi-\arcsin(y)$ for $y > 0$
  • $x = \pi-\arcsin(y)$ and $2\pi + \arcsin(y)$ for $y < 0$.

Hence, $Y$ has positive values if $X$ takes values in $A_1=(0,\arcsin y)$ or $A_2=(\pi-\arcsin y,2\pi)$; $Y$ has negative values if $X$ takes values in $B=(\pi-\arcsin y,2\pi + \arcsin y)$. enter image description here

So we have for $-1<y<1$, $$ \Bbb P(Y\le y)=\begin{cases} \Bbb P(X\in B)& -1<y< 0\\ \Bbb P(X\in A_1\cup A_2)=\Bbb P(X\in A_1)+\Bbb P(X\in A_2)& 0<y< 1\\ \end{cases} $$ that is $$ F_Y(y)=\begin{cases} F_X(2\pi + \arcsin y) -F_X(\pi-\arcsin y)& -1<y< 0\\ F_X(\arcsin y) +1-F_X(\pi-\arcsin y)& 0<y< 1\\ \end{cases} $$ and then $$ F_Y(y)=\frac{\pi + 2 \arcsin(y)}{2\pi},\qquad y\in (-1,1) $$

The distribution function of $Y$ is $$ F_Y(y)=\begin{cases} 0 & y\le 0\\ \frac{\pi + 2 \arcsin(y)}{2\pi} & y\in (-1,1)\\ 1 & y\ge 1 \end{cases} $$ We differentiate the above expression to obtain the probability density: $$ f_Y(y)=\begin{cases} \frac{1}{\pi\sqrt{1-y^2}} & y\in (-1,1)\\ 0 & y\notin (-1,1) \end{cases} $$

  • $\begingroup$ I follow you when you say there are two solutions on the interval for the equation. I do not understand how you come up $F_y(y)$. Please explain it. $\endgroup$ – Bob Mar 25 '16 at 14:18
  • $\begingroup$ I've added some details. $\endgroup$ – alexjo Mar 25 '16 at 15:36

your answer is correct indeed just a little mistake. in the first step: $$P(Y \le y) = P( \sin x \le y ) \Rightarrow P(x_1 \le x \le x_2 )$$ as shown in figure below (sorry figure is for $y=sin(x+\theta)$ but still is useful .just draw the figure for $sin(x)$ in your imagination. also figure is from "Probability, Random Variables and Stochastic Processes" by Papoulis A. chapter 5 example 7. which is very similar to your problem). enter image description here

since $sin(x)$ intersects with each horizontal line two times. so we can now write the right answer as below: $$P(Y \le y)= \frac{1}{2\pi} \int_0^{x_{1}}dx+\frac{1}{2\pi} \int_{x_{2}}^{2\pi}dx$$ and we have in mind that $x_2=\pi-\arcsin(y)=\pi-x_1$ so we get:

$$P(Y \le y)= \frac{1}{2\pi} \int_0^{x_{1}}dx+\frac{1}{2\pi} \int_{x_{2}}^{2\pi}dx = \frac{1}{2\pi}\times (x_{1}+2\pi-x_{2})=\frac{1}{2\pi}\times (x_{1}+\pi+x_{1}) = \frac{1}{2\pi}\times (2x_{1}+\pi)=\frac{1}{2}+\frac{1}{\pi}\arcsin(y)$$

now by taking derivative with respect to $y$ we get: $$f(y) = \frac{1}{\pi \sqrt{1 - y^2}}$$


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.