Name

dfp_min — DFP minimization

CALLING SEQUENCE

[result]=dfp_min(func,pin[,arg1,…argn])

PARAMETERS

Input

• func = (minus likelihood) function to minimize

• pin = parameter vector fed to func

• argi =

 - 'maxit=x' to set the maximum # of iterations (default = 500)

 - arguments of function func others that the parameters

Output

• a result list with:

  - result('meth') = 'dfp'

  - result('b') = (kx1) minimizing vector

  - result('f') = value of func at solution values

  - result('hess') = hessian at solution values

  - result('iter') = # number of iterations

  - result('flag') = 0 for convergence, 1 for non-convergence

DESCRIPTION

DFP minimization routine to minimize func (Converted from Numerical Recipes book dfpmin routine). func must take the form func(b,varargin) where: b = parameter vector (k x 1) varargin = arguments passed to the function

EXAMPLE

1) res = ols2(y,x);
2) result4 = dfp_min('to_llike', [res('beta'); res('sige')],'maxit=1000',y,x);
 
This example, taken from optim1_d, realises the estimation of a tobit model, with initial values taken from on ols estimation. Maximum # of iterations is set to 1000.  
 

               

AUTHOR

Eric Dubois 2002