optimize to maximize the
conditional expectation log likelihood in alpha. The interval is
set to [alpha_old/2, 2*alpha_old], supposing that the previous guess of
alpha_old is not far from reality.
estimate.alpha(phylo, conditional_law_X, sigma2, mu, shifts, alpha_old,
max_selection.strength)Input tree.
result of function compute_E.OU
variance of params
mean of the root state
list of shifts on the tree
previous estimation of the selection strength
the maximal value of alpha authorized by the user
double : estimation of alpha
This function uses functions compute_var_diff.OU
and compute_diff_exp.OU in the process. Careful : only works if the
root is stationary, and shifts at nodes.