1D Gaussian

Learn More

1D Gaussian

Interactive visualization of the univariate Gaussian density

$$ f(x) = \frac{1}{\sqrt{2\pi}\sigma} \cdot \exp\!\left\{ -\frac{1}{2} \cdot \left(\frac{x-\mu}{\sigma}\right)^2 \right\}\enspace, \quad x\in \mathbb{R} \enspace, $$

with mean $\mu \in \mathbb{R}$ and standard deviation $\sigma \in \mathbb{R}_+$ .

It was discovered by Karl Friedrich Gauß (1777-1855) in Göttingen, Germany.

Formulas

  • quantile function
    $Q(p) = \mu + \sigma\, \sqrt{2}\, \mathrm{erf}^{-1}(2p-1)$
  • uniform to Gaussian
    $x_i^{\text{Gauss}} = Q(x_i^{\text{uni}})$
  • golden Kronecker sequence
    $x_i^{\text{uni}}=\mod( \Phi \cdot (i+z), 1) \enspace, \quad i \in \{1,2,\ldots,L\}\enspace, \quad z \in \mathbb{Z}$
  • equidistant samples
    $x_i^{\text{uni}} = \frac{2 i - 1 + \gamma}{2 L} \enspace, \quad i \in \{1,2,\ldots,L\}\enspace,\quad \gamma\in[-1,1]$
  • unscented transform [Julier97], $L=3$, parameter $\kappa > -1$
    $x_0=\mu \enspace, \quad W_0 = \frac{\kappa}{1+\kappa} \enspace, \quad x_{1,2}=\mu \mp \sqrt{1+\kappa}\,\sigma \enspace, \quad W_{1,2} = \frac{1}{2(1+\kappa)}$
    $\kappa=0$ yields $L=2$ samples $\mu\pm\sigma$, $\kappa=2$ additionally matches the 4th moment (kurtosis) of the Gaussian. Sample lines are drawn with height proportional to their weight, the largest weight reaching the density maximum.