Understanding steps for integrating standard normal distribution using polar coordinates

217 Views Asked by At

I have the following derivation for showing that the integral of the standard normal distribution is equal to one.

I don't completely understand the steps so I'll provide the derivation and then write what I think is happening.

$$ \begin{align} 2\pi \int\limits_{-\infty}^{\infty} p(x) dx &= 2\pi \int\limits_{-\infty}^{\infty} p(x) dx \tag{1} \\ &= 2 \pi \int\limits_{-\infty}^{\infty} p(x) dx \int\limits_{-\infty}^{\infty} p(y) dy \tag{2}\\ &= \int\limits_{-\infty}^\infty\int\limits_{-\infty}^\infty e^{-(x^2+y^2)/2}dx\ dy \tag{3}\\ &= \int\limits_{\theta=0}^{2\pi} \int\limits_{r=0}^\infty e^{-r^2/2} r\ dr\ d\theta \tag{4} \\ &= 2\pi \tag{5} \end{align} $$

  1. Is what we're trying to integrate (wasn't included in the text but I added it for clarity). I think both sides should be multiplied by $2\pi$ which we do to make the leap in (3) easier so I included that.

  2. We're multiplying by the integral in terms of $y$. I think the argument is that this is similar to multiplying by 1, but I'm not fully satisfied by this because we're supposed to be proving that. I think this fallacy is called "begging the question".

  3. Integration is commutative. I'm comfortable with the inside argument but I'm not sure why the $2\pi $ got consumed.

  4. We converted to polar coordinates. No problem here.

  5. Evaluate the integral. No problem here.

So as was explained earlier I'm worried about begging the question in (2) and I don't know why the $\boldsymbol{2\pi}$ got consumed in (3).

1

There are 1 best solutions below

0
On BEST ANSWER

Let $I$ be the value of the integral. Then get rid of line (1), and on the left hand side of equation (2) write $2\pi I^2 = ...$.

You end up with $2\pi I^2 = 2\pi$. Then solve for $I$, using the fact that $I \ge 0$.