MHB Recovering PMF from characteristic equation

Click For Summary
The discussion focuses on the integration of the characteristic equation to recover the probability mass function (PMF), leading to unexpected results, specifically a value of zero. The calculations reveal that for integer values of k, the right-hand side becomes zero due to the sine function, which complicates the definition of the PMF. Additionally, the expression for the PMF, \(p_X[k] = \frac{-2k\sin(\pi k)}{k^2 - 1}\), is problematic as it is undefined for k equal to -1 or 1. The participants are prompted to investigate the limits as k approaches these values to identify the source of the issue. The discussion highlights the challenges in correctly defining the PMF from the characteristic equation.
Dustinsfl
Messages
2,217
Reaction score
5
I am integrating the characteristic equation in order to recover the PMF, but I am going to get the answer to be zero so something went wrong.
\begin{align*}
p_X[k]
&= \frac{1}{2\pi}\int_{-\pi}^{\pi}\phi_X(\omega)e^{-i\omega k}d\omega\\
&= \frac{1}{2\pi}\int_{-\pi}^{\pi}\cos(\omega)e^{-i\omega k}d\omega\\
&= \frac{i}{2\pi k}\cos(\omega)e^{-i\omega k}\bigg|_{-\pi}^{\pi} +
\frac{i}{2\pi k}\int_{-\pi}^{\pi}\sin(\omega)e^{-i\omega k}d\omega\\
&= \frac{i}{2\pi k}(e^{i\pi k} - e^{-i\pi k}) + \frac{i}{2\pi k}\Bigg[
\frac{i}{k}\sin(\omega)e^{-i\omega k}\bigg|_{-\pi}^{\pi} -
\frac{i}{k}\int_{-\pi}^{\pi}\cos(\omega)e^{-i\omega k}d\omega\Bigg]\\
&= \frac{-1}{\pi k}\bigg(\frac{e^{i\pi k} - e^{-i\pi k}}{2i}\bigg) +
\frac{1}{2\pi k^2}\int_{-\pi}^{\pi}\cos(\omega)e^{-i\omega k}d\omega\\
\bigg(\frac{k^2 - 1}{2\pi k^2}\bigg)
\int_{-\pi}^{\pi}\cos(\omega)e^{-i\omega k}d\omega
&= \frac{-\sin(\pi k)}{\pi k}\\
\int_{-\pi}^{\pi}\cos(\omega)e^{-i\omega k}d\omega
&= \frac{-2k\sin(\pi k)}{k^2 - 1}
\end{align*}
Since k is an integer, the RHS is zero.
Additionally, the \(p_X[k]\) is supposed to be defined for all integers, but if I say \(p_X[k] = \frac{-2k\sin(\pi k)}{k^2 - 1}\), k cannot be -1 or 1. This definition of the PMF would be weird anyways since \(\sin(\pi k) = 0\) anyways.

What went wrong?
 
Last edited:
Physics news on Phys.org
You can compute the limits $k \to 1$ and $k \to -1$.