lambda.reg(object, columns)eiReg, the output from ei.reglambda.regmsmmsm. The arguments passed to
deltamethod in msm include
ga list of transformations of the form ~ x1 / (x1 + x2 +
+ ... + xk), ~ x2 / (x1 + x2 + ... + xk), etc.. Each
$xc$ is the estimated proportion of all row members in column
$c$, $beta_rc$
meanthe estimated proportions of the row members in the
specified columns, as a proportion of the total number of row
members, $(beta_r1, beta_r2, ..., beta_rk)$.
cova diagonal matrix with the estimated variance of each
$beta_rc$ on the diagonal. Each column
marginal is assumed to be independent, such that the off-diagonal
elements of this matrix are zero. Estimates come from
object$cov.matrices, the estimated covariance matrix from
the regression of the relevant column. Thus,
| cov | = | $Var(beta_r1)$ | 0 | 0 |
| $...$ | 0 | $Var(beta_r2)$ | ||
| 0 | $...$ | 0 | ||
| 0 | $Var(beta_{r3})$ | $...$ | ||
| $...$ | $...$ | $...$ | $...$ | cov |
ei.reg