Gaussian concentration

For X𝒩(μ,σ2)X \sim \mathcal{N}(\mu,\sigma^2): Pr[X=μ±x]1σ2πex2/2σ2\mathrm{Pr}[X=\mu \pm x] \sim \frac{1}{\sigma\sqrt{2\pi}}e^{-x^2/2\sigma^2}

See Gaussian tail bound


Gaussian function

Gaussian function takes the form f(x)=exp(x2)f(x) = \exp(-x^2), and parametric extension f(x)=aexp((xb)22c2)f(x)=a\exp \left(-{\frac {(x-b)^{2}}{2c^{2}}}\right), a,b,ca,b,c \in \mathbb{R}.

Probability density function of normally distributed random variable, expected value μ=b\mu = b, variance σ2=c2\sigma^2 = c^2, g(x)=1σ2πexp(12(xμ)2σ2)g(x)={\frac {1}{\sigma {\sqrt {2\pi }}}}\exp \left(-{\frac {1}{2}}{\frac {(x-\mu )^{2}}{\sigma ^{2}}}\right)


See also: Concentration inequality


References:

  1. Gaussian function - Wikipedia: