what is the upper bound for the normal cdf

23 Views Asked by At

In wikipidia of error function: https://en.wikipedia.org/wiki/Error_function, there are few words saying that

The error function and its approximations can be used to estimate results that hold with high probability or with low probability. Given a random variable X ~ Norm[μ,σ] (a normal distribution with mean μ and standard deviation σ) and a constant L > μ, it can be shown via integration by substitution:

\begin{aligned} \operatorname{Pr}[X \leq L] & =\frac{1}{2}+\frac{1}{2} \operatorname{erf} \frac{L-\mu}{\sqrt{2} \sigma} \\ & \approx A \exp \left(-B\left(\frac{L-\mu}{\sigma}\right)^2\right) \end{aligned}

where A and B are certain numeric constants. If L is sufficiently far from the mean, specifically μ − L ≥ σ√ln k, then:

$$\operatorname{Pr}[X \leq L] \leq A \exp (-B \ln k)=\frac{A}{k^B}$$.

My question is, how to determine A and B such that the above inequality always holds?