1
$\begingroup$

How do I go about finding the the pdf of the statisitc $\sum_i x_i ^2$ such that each $x_i$ is iid from a $N(\sigma , \sigma)$ distribution? I've searched, but cannot find a straightforward answer. I've looked at charts such as this (which I highly recommend reviewing if you have not seen one like it). I've also reviewed this answer about the noncentral chi-squared distribution, but it seems to be specifically for iid variables drawn from the distribution $N(0, \sigma^2)$.

$\endgroup$
  • $\begingroup$ Just to be clear: $X\sim N(\sigma,\sigma)$ means that $X$ has mean $\sigma$ and variance $\sigma$ as usual? $\endgroup$ – Jack D'Aurizio May 6 '15 at 14:55
  • $\begingroup$ @JackD'Aurizio yes that is the case. $\endgroup$ – kathystehl May 6 '15 at 15:51
1
$\begingroup$

If $X_i$ are iid normal with mean $\mu$ and variance $\sigma^2$, then $\sigma^{-2} \sum_{i=1}^n X_i^2$ has noncentral chi-squared distribution with $n$ degrees of freedom and noncentrality parameter $\lambda = n (\mu/\sigma)^2$ (as defined e.g. in Wikipedia; conventions may sometimes differ). So what you have is a multiple of a noncentral chi-squared random variable. If $Y$ has pdf $f_Y$, then $Z = c Y$ (for $c > 0$) has pdf $f_Z(z) = c^{-1} f_Y(z/c)$.

$\endgroup$
  • $\begingroup$ Thank you @Robert. To clarify, what would $c$ be in the example above? $\endgroup$ – kathystehl May 6 '15 at 15:54
  • $\begingroup$ Let me use $a$ instead of your $\sigma$, to avoid confusion. If $\mu = \sigma^2 = a$, then $Y = \sigma^{-2} \sum_{i=1}^n X_i^2$ has the noncentral chi-square distribution with $\lambda = n a$, and your statistic $Z = \sum_{i=1}^n X_i^2 = \sigma^2 Y = a Y$. $\endgroup$ – Robert Israel May 7 '15 at 2:44

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.