# matrix dimensions do not agree?

2 views (last 30 days)
john birt on 26 Apr 2012
how can I add more observed variables into a function 'f'
x = [1.4 1.5 1.6]; z=[4.1 5.3 6.7]; y = [0.2 0.31 0.43];
c = 0.1; %initial guess
t=0.1:0.1:1;
f = @(c,x) arrayfun(@(x)trapz(t,x.*c.*t.^2),x);
cfit = nlinfit(x,y,f,c)
works well, but trying to adapt the function to use more observed information in the form of the vector 'z' by coding
f = @(c,x) arrayfun(@(x)trapz(t,z+x.*c.*t.^2),x);
gives an error
??? Error using ==> nlinfit at 120
Error evaluating model function
'@(c,x)arrayfun(@(x)trapz(t,z+x.*c.*t.^2),x)'.
Error in ==> test88 at 6
cfit = nlinfit(x,y,f,c)
Caused by:
Error using ==> plus
Matrix dimensions must agree.

Andrei Bobrov on 26 Apr 2012
x = [1.4 1.5 1.6]; z=[4.1 5.3 6.7]; y = [0.2 0.31 0.43];
c = 0.1; %initial guess
t=0.1:0.1:1;
f = @(c,x) arrayfun(@(x,z)trapz(t,x.*c.*t.^2),x,z);
cfit = nlinfit(x,y,f,c)
##### 2 CommentsShowHide 1 older comment
john birt on 26 Apr 2012
it works!!!
clear;
x = [1.4 1.5 1.6]; z=[4.1 5.3 6.7]; y = [0.2 0.31 0.43]; n=size(x);
c = 0.1; %initial guess
t=0.1:0.1:1;
f = @(c,x) arrayfun(@(x,z)trapz(t,z+x.*c.*t.^2),x,z);
[cfit, r] = nlinfit(x,y,f,c)
error = sum(r.^2)/n(2)