init.alpha.estimation fits (Y_i-Y_j)^2 ~ gamma^2(1-exp(-alpha*d_ij))
for all couples of tips (i,j) that have the same mean, i.e than are not
separated by a shift. Shifts are initialized thanks to a lasso
(function init.EM.lasso).
init.alpha.gamma.estimation(
phylo,
Y_data,
nbr_of_shifts,
times_shared,
distances_phylo,
T_tree,
subtree.list,
max_triplet_number,
alpha_known,
method.init.alpha.estimation,
tol_EM,
h_tree,
miss,
masque_data,
independent,
...
)params_init the list of initial parameters to be used, in the right format.
a phylogenetic tree, class phylo.
data at the tips.
number of shifts used in the EM algorithm
(matrix) : phylogenetic distance, result of function
compute_dist_phy
Function robustbase::nlrob is used for the robust fit.