324 Boolean Analysis — Threshold Functions — Gaussian Polar
324.1 Overview
Planar Gaussian density separates into radial and angular factors. Cone masses and the angular overlap of two halfplanes give the geometric formula behind Gaussian quadrant probabilities.
324.2 Declarations
For \(z=(x,y)\in \mathbb R^2\), the planar Gaussian density is \(d_2(x,y)=(2\pi )^{-1}\exp (-(x^2+y^2)/2)\).
For all real \(x,y\), let \(\varphi (t)=(2\pi )^{-1/2}e^{-t^2/2}\) and \(d_2(x,y)=(2\pi )^{-1}e^{-(x^2+y^2)/2}\). Then \(\varphi (x)\varphi (y)=d_2(x,y)\), with both sides regarded as nonnegative extended real values.
Let \(\gamma \) be the standard real Gaussian probability measure. For every measurable \(A\subseteq \mathbb R^2\), \((\gamma \otimes \gamma )(A)=\int _A(2\pi )^{-1}e^{-(x^2+y^2)/2}\, dx\, dy\), where the integral is nonnegative and extended-valued.
Let \(f(r)=r\exp (-r^2/2)\) for \(r\gt 0\). Its nonnegative extended integral satisfies \(\int _{(0,\infty )}f(r)\, dr=1\).
Let \(A\subseteq \mathbb R^2\) and \(\Theta \subseteq \mathbb R\) be measurable sets such that for all real \(r\gt 0\) and \(\theta \), \((r\cos \theta ,r\sin \theta )\in A\) if and only if \(\theta \in \Theta \). Then \(\int _A(2\pi )^{-1}e^{-(x^2+y^2)/2}\, dx\, dy=(2\pi )^{-1}\operatorname {Leb}(\Theta \cap (-\pi ,\pi ))\), with the integral and angular Lebesgue measure in the nonnegative extended reals.
For every real \(\theta \) with \(-\pi \lt \theta \lt \pi \), \(\cos \theta \le 0\) if and only if \(\theta \le -\pi /2\) or \(\pi /2\le \theta \).
For every real \(u\) with \(-2\pi \lt u\le -\pi /2\), \(\cos u\le 0\) if and only if \(-3\pi /2\le u\).
For every real \(u\) with \(-\pi /2\le u\lt \pi \), \(\cos u\le 0\) if and only if \(\pi /2\le u\) or \(u=-\pi /2\).
For every real \(\alpha \), the set \(\{ \theta \in \mathbb R:\cos \theta \le 0\ \text{and}\ \cos (\theta -\alpha )\le 0\} \) is Borel measurable.
For every real \(\alpha \in [0,\pi ]\), the Lebesgue measure of \(\{ \theta \in (-\pi ,\pi ):\cos \theta \le 0,\ \cos (\theta -\alpha )\le 0\} \) equals \(\pi -\alpha \).