... ... @@ -128,17 +128,6 @@ Similarly the informative counter-part of $VarLs(X)$ is defined as $\ivarls(X)=\ \begin{methods} \section{Methods} Two methods are proposed to estimate$\overline{\rcovls(X,Y)}$. The first one is formal and applicable only when$p=1$and$q=1$the second one is based on a Monte-Carlo evaluation and is applicable for every$p$and$q$. \subsection{Formal estimation of$\overline{\rcovls(X,Y)}$} For two random real vectors of length$n$,$x$and$y$the mean of$R(x,y)^2$,$\overline{\r(x,y)^2}=1/(n-1)$. That equality is independent of the distribution of$x$and$y$. Let$\sigma_x$and$\sigma_y$being respectively the standard deviations of$x$and$y$, we can estimate$\overline{\rcovls(x,y)}$by Equation~(\ref{eq:RCovLs11}). \begin{equation} \overline{\rcovls(x,y)} = \sigma_x \; \sigma_y \sqrt{\frac{1}{n-1}} \label{eq:RCovLs11} \end{equation} \subsection{Monte-Carlo estimation of$\overline{\rcovls(X,Y)}$} For every values of$p$and$q$including$1$,$\overline{\rcovls(X,Y)}$can be estimated using a serie of$k$random matrices$RX=\{RX_1,RX_2,...,RX_k\}$and$RY=\{RY_1,RY_2,...,RY_k\}$where each$RX_i$and$RY_i$have the same structure respectively than$X$and$Y\$ in term of number of columns and of standard deviation of these columns. ... ...