I found some scholars clarified that the relationship between the Shannon transform and Stieltjes transform with

\begin{equation}

\frac{\gamma}{\log e}\frac{d}{d\gamma}\mathcal{V}_N(\gamma) = 1 - \frac{1}{\gamma}\mathcal{S}_N(-\frac{1}{\gamma})

\end{equation}

However, other scholars claimed that this equality was modified as

\begin{equation}

\mathcal{V}_N(x) = \frac{1}{N}\log \det (\mathbf{I}_N + \frac{1}{x}\mathbf{B}_N) = \int_{0}^{+\infty}\log (1+\frac{\lambda}{x})dF_N(\lambda) = \int_{x}^{+\infty}(\frac{1}{\omega}-\mathcal{S}_N(-\omega))d\omega

\end{equation}

Here $\mathcal{S}_X(.)$ denotes the stieltjes transform, whereas $\mathcal{V}_X(.)$ is the shannon transform. So how can I obtain the last equation from the first equation?

***I mean, given that***

\begin{equation}

\frac{\gamma}{\log e}\frac{d}{d\gamma}\mathcal{V}_N(\gamma) = 1 - \frac{1}{\gamma}\mathcal{S}_N(-\frac{1}{\gamma})

\end{equation}

***How can I obtain this***

\begin{equation}

\mathcal{V}_N(x) = \int_{x}^{+\infty}(\frac{1}{\omega}-\mathcal{S}_N(-\omega))d\omega

\end{equation}

I did some calculations, but not sure that they are equal.

More Di Zhang's questions See All
Similar questions and discussions