4
$\begingroup$

Let $M_n$ be the matrix $$M_n=\begin{pmatrix} 1&\binom{1}{1}\binom{1-1}{1-1} &0 &0\qquad \qquad \dots &0\\ 1&\binom{2}{1}\binom{2-1}{1-1} &\binom{2}{2}\binom{2-1}{2-1} &0 \qquad\qquad\dots & 0\\ 1&\binom{3}{1}\binom{3-1}{1-1} &\binom{3}{2}\binom{3-1}{2-1} &\binom{3}{3}\binom{3-1}{3-1}\qquad\dots &0\\ \qquad \qquad \dots\\ \qquad \qquad \dots\\ 1&\binom{n}{1}\binom{n-1}{1-1} &\binom{n}{2}\binom{n-1}{2-1} &\binom{n}{3}\binom{n-1}{3-1}\qquad\dots &\binom{n}{n-1}\binom{n-1}{n-2}\end{pmatrix}.$$

Question. It appears that the modified Bessel function of the first kind, of order $0$, has the expansion $$I_0(x)=\exp\left(\sum_{j=1}^{\infty}\frac{(-1)^{j-1}\det(M_j)}{j!^2}\left(\frac{x}2\right)^{2j}\right).$$ Is this true?

$\endgroup$
1
  • 2
    $\begingroup$ The Bessel function $I_0$ is the moment generating function of the variable with PDF ~ $1/\sqrt{1-x^2}$. Essentially you are asking for the cumulants. The cumulants can be obtained from the moments. The moments in turn are given by the Taylor series of $I_0$ which is simple, i.e. $m_{2 k } = 2^{-2k}/k!$ (I would say..). $\endgroup$
    – lcv
    Commented Apr 14, 2017 at 1:12

0

You must log in to answer this question.