explodes a Varmax model for estimation uses
[theta,theta2mat,V2theta,thetalab,AR,ARS,MA,MAS,G,V,p,P,q,Q,g,s,k,n,np,%type,vdiag,ineq]=arma2param(m,AR,ARS,MA,MAS,V,s,G,r,namexos)
* m = # of endogenous variables
* AR = a string, representing the name of the AR part of the model
* ARS = a string, representing the name of the seasonal AR part of the model
* MA = a string, representing the name of the MA part of the model
* MAS = a string, representing the name of the seasonal MA part of the model
the names of FR, FS, AR and AS represent objects that can be either of constant type or list type; in the last case, the first item of the list should represent as in the fist case the starting value of the corresponding matrix, and second one a string matrix, of same size, of constraints ('' for no constraint, '=' for equality constraint, 'value< *', '*< value' or 'value1< *< value2' for inequality constraints)
* V = a (mxm) (var-covar) matrix or a (mx1) vector (the diagonal of a var-covar matrix, supposed to have 0 outside the diagonal)
* s = order of seasonality
* G = the coefficients martrix for the endogenous variables (if any)
* r = # of exogenous variables (if any)
* theta = (npx1) vector of parameters that will be estimated
* theta2mat = vector of strings, representing the instructions that transform back theta into the input parameters
* V2theta = vector of strings, representing the instructions that transform V into the corresponding parameters in theta thetalab = (npx1) vector of strings, representing theta names used when printing the results
* FR = the AR part of the model
* FS = the seasonal AR part of the model
* AR = the MA part of the model
* AS = the seasonal MA part of the model
* G = the coefficients matrix for the endogenous variables
* V = the (mxm) (var-covar) matrix
* p = degree of the AR part of the model
* P = degree of the seasonal AR part of the model
* q = degree of the MA part of the model
* Q = degree of the seasonal MA part of the model
* s = order of seasonality
* k = maximum degree of the total AR, MA and G parts
* n = k*m
* np = # of estimated parameters
* %type = type of the e4 model
* vdiag = 0 if V is diagonal, 1 if not