[Scilab-users] Command datafit for a bivarate model function z=f(x, y)
Jens
j.s.strom at hslmg.de
Mon Dec 15 23:00:43 CET 2014
I tried to follow the pattern for data fit given in the thread
http://mailinglists.scilab.org/Convert-x-y-z-data-into-a-z-f-x-y-function-td4026897.html#a4026925
for the simple bivariate function
z=p(1)*x.^2 + p(2)*y.^2
by this script:
mode(0), lines(0), clc(), clear
deff( 'z=!z(x,y,p)','z=p(1)*x.^2 + p(2)*y.^2')//model function
pg=[5;6]//parameter of fictitious data
X=[0:5];Y=[0:5]; Z=!z(X,Y,pg)
M=[X;Y;Z];//measurement matrix, 3 rows, 6 colomns
deff(' e=!e(p,m)','x=M(1), y=M(2); z=M(3),e=z-!z(x,y)')//defect
(critereon- , error-,) function
p0=[2;3];
[p,err]=datafit(1,!e,M,p0)
err,
z=!z(X,Y,p)
However the script just returns p0 as solution. Would anyone kindly help me
to find the flaw?
Kind regards
Jens
--
View this message in context: http://mailinglists.scilab.org/Command-datafit-for-a-bivarate-model-function-z-f-x-y-tp4031533.html
Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com.
More information about the users
mailing list