


lyngby_lzit_main - Main function for Lange-Zeger estimation
function Result = lyngby_lzit_main(P, Y, 'PropertyName',
'PropertyValue');
Input: P Paradigm
Y Normalized data matrix
Property: Iterations { 90 } The number iterations
MinChange { 1e-4 } Stopcriterion for the
iteration
StepSize { 1 } Gradient descent stepsize
Theta1Init { 10 } Initialization of theta1.
Theta2Init { 2 } Initialization of theta2.
Output: Result Estimated parameters in matrix 3xN
[beta; theta1; theta2]'
Fitting of a gamma kernel for each voxel in the Fourier domain.
This function makes some initialization before iterating:
for every voxel calling lyngby_lzit_algo to do the actual fitting.
Ref: Lange, Zeger (1997), Roy Stat Soc, Appl Stat. 44.
See also LYNGBY, LYNGBY_LZGS_MAIN, LYNGBY_LZIT_ALGO,
LYNGBY_LZITPICKF, LYNGBY_LZIT_FTLAMB,
LYNGBY_LZIT_THOPT, LYNGBY_LZIT_NOISE,
LYNGBY_LZIT_BETA.
$Id: lyngby_lzit_main.m,v 1.8 2003/02/12 20:06:49 fnielsen Exp $