1
$\begingroup$

I have computed a truncated normal distribution, which total probability density (i.e. area under the curve) is equal to 0.92.

The distribution represents well the reality of the phenomenon I am investigating, i.e. I am expecting that when x = 0 the probability will be positive (non-zero). The mode is also where I expected it to be. However, any non represented probabilities for x < 0 do not make physical sense, thus I would like to "rescale" my truncated distribution so that the total area under the density is 1 (not 0.92).

The distribution being asymetrical, how can I do this?

$\endgroup$
1
  • $\begingroup$ There's no point in a truncated normal where there's positive probability for that point. Either you don't mean 'probability' or you don't mean 'truncated'. $\endgroup$ – Glen_b Aug 5 '14 at 12:01
2
$\begingroup$

Divide the constant term out the front by the area. That rescaled density has area 1.

So if the area under $c_0 g_X(x)$* over the truncated domain is exactly 0.92, then the area under $\frac{c_0}{0.92}g_X(x)$ on the same domain is 1.

*(where $g$ is some function normalized by $c_0$ to integrate to 1 on the untruncated domain)

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.