TreeSelection(y, yh, z, zh, group, zo3, x0, ShowInfo)

Generalized Heckman model estimation

Estimates generalized Heckman model (theoretically developed by B.Potanin and E.Kossova). Function returns [twostep,MLE] where twostep are two-step estimations while MLE are maximum-likelihood estimations.
y is dependend variable.

yh are independend variables explaining y.

z is a cell matrix of selection equation independend variable

z is a cell matrix of selection equation dependend variables

group says which group does the observation belongs to. Groups differ in a error term distribution. Also, special 0 group undermines that we have no observations. So in classical Heckman method the groups are 0 and 1.

zo3 describes selection variable outcome for a given group.

x0 if proviede gives initial points