help with inner dimention mismatch

Hi,

I am trying to run the following code and get an error(given). Could someone please help me figure it out?

Thanks.

clc;
clear;

p=1;
T = 1000;
h=10;
B=1000;
R=1000;
toyo=[];

simulations = 1000;

ftyPI = zeros(h,simulations);
fsigPI = zeros(h,simulations);

Lbyth = zeros(h,simulations);
Lbsigh = zeros(h,simulations);

mccounty = zeros(h,simulations);
mccountsig = zeros(h,simulations);

lb = eps*[1 0]';

a0 = 0.1; a1 = 0.4; 

ra = zeros(T+2000,1);
seed = 54321;
rng(seed);

for mc= 1:simulations

    mc
    %ra = randn(T+2000,1); 
    ra = trnd(5,T+2000,1);
    ra = ra./(sqrt(5/3));

       ytn = [];

       yt1=zeros(T/5,1);
       yt2=zeros(T/5,1);
       yt3=zeros(T/5,1);
       yt4=zeros(T/5,1);
       yt5=zeros(T/5,1);

epsi=zeros(T+2000,1);
simsig=zeros(T+2000,1);

unvar = a0/(1-a1);

for i = 1:T+2000
                 if (i==1) 
                     simsig(i) = a0+a1*((a0)/(1-a1));
                     s=(simsig(i))^0.5;
                     epsi(i) = ra(i) * s;

                   else                     
                       simsig(i) = a0+ a1*(epsi(i-1))^2;
                       s=(simsig(i))^0.5;
                       epsi(i) = ra(i)* s;
                   end
  end

yinteta=epsi(2001:T+2000);

epsi2 = epsi.^2;
yt = epsi2(2001:T+2000);

bytlast = yt(T);
tytlast = yt(T);

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%ESTIMATION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

yint = epsi(2001:T+2000);

ytop = yte(yint,T);

theta0 = 0.5*ones(7,1)';
Av=[0 1 0 0 0 0 0 ];
bv=0.99999;

opts = optimset('Display','off','Algorithm','interior-point');
[theta, opt] = fmincon(@(theta)...
               lhnew(theta,ytop),theta0,Av,bv,[],[],lb,[],[],opts);

alpha0 = theta(1);
ar1= theta(2);
cc1 = theta(3); 
cc2 = theta(4); 
cc3 = theta(5); 
cc4 = theta(6); 
cc5 = theta(7); 

mali = [alpha0 ar1 cc1 cc2 cc3 cc4  cc5];

toyo=[toyo;mali];

end

mean(toyo)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

function ytop = yte(yint,T)

yin = yint;

for i=1:T
    m = mod(i,5);
    if  m==0
        m=5;
    end  

      if m==1
      yin(i)= yin(i)*(2^0.5);
      elseif m==5
      yin(i)= yin(i)*1;
      else
      yin(i)= yin(i)*(0.5^0.5);   
      end
  end 

ytop=yin;

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

function L = lhnew(theta,y)

w     =  theta(1); 
alpha =  theta(2);
c1    =  theta(3); 
c2    =  theta(4);
c3    =  theta(5); 
c4    =  theta(6); 
c5    =  theta(7); 

y2   = y.^2;
[T1,K] = size(y2); 
ht    = zeros(T1,1);
ht(1) = sum(y2);

for i=2:T1

      m = mod(i-1,5);
      if  m==0
          m=5;
      end

      if m==1
      ht(i)= w + alpha*c1*y2(i-1);
      elseif m==2
      ht(i)= w + alpha*c2*y2(i-1);
      elseif m==3
      ht(i)= w + alpha*c3*y2(i-1);
      elseif m==4
      ht(i)= w + alpha*c4*y2(i-1);
      elseif m==5
      ht(i)= w + alpha*c5*y2(i-1);
      end
  end

l=zeros(T1,1);

for i=1:T1

      m = mod(i,5);
      if  m==0
          m=5;
      end

      if m==1
      l(i) = -0.5*log(2*pi) - log(sqrt(ht(i)) - 0.5*((c1*y(i)/sqrt(i))^2));
      elseif m==2
      l(i) = -0.5*log(2*pi) - log(sqrt(ht(i)) - 0.5*((c2*y(i)/sqrt(i))^2));
      elseif m==3
      l(i) = -0.5*log(2*pi) - log(sqrt(ht(i)) - 0.5*((c3*y(i)/sqrt(i))^2));
      elseif m==4
      l(i) = -0.5*log(2*pi) - log(sqrt(ht(i)) - 0.5*((c4*y(i)/sqrt(i))^2));
      elseif m==5
      l(i) = -0.5*log(2*pi) - log(sqrt(ht(i)) - 0.5*((c5*y(i)/sqrt(i))^2));
      end
  end

l=-l;
L = sum(l);

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
error:

Error using  * 
Inner matrix dimensions must agree.

Error in C:\Program Files\Matlab\R2012a\toolbox\optim\optim\private\backsolveSys.p>backsolveSys (line 18)

Error in C:\Program Files\Matlab\R2012a\toolbox\optim\optim\private\solveAugSystem.p>solveAugSystem (line 23)

Error in C:\Program Files\Matlab\R2012a\toolbox\optim\optim\private\leastSquaresLagrangeMults.p>leastSquaresLagrangeMults (line 28)

Error in C:\Program Files\Matlab\R2012a\toolbox\optim\optim\barrier.p>barrier (line 626)

Error in fmincon (line 841)
    [X,FVAL,EXITFLAG,OUTPUT,LAMBDA,GRAD,HESSIAN] = barrier(funfcn,X,A,B,Aeq,Beq,l,u,confcn,options.HessFcn, ...

Error in test (line 203)
[theta, opt] = fmincon(@(theta)...

Answer 1

[Image Analyst]

I didn't run the code but in general what it means is that if you're doing a matrix multiplication of a r1 by c1 matrix by a r2 by c2 matrix (that's rows by columns), then c1 must equal r2:

(r1 x c1) * (r2 x c2) must = (r1 x c1) * (c1 x c2) to give (r1 x c2). For some reason, you're not using matrices with those dimensions.

[Image Analyst]

The "inner dimensions" refer to the c1 and r2 - those must be equal, and they sort of cancel out and you're left with a matrix that has the outer dimensions of r1 x c2. At least that's how I think of it (the inner numbers vanishing or canceling out: m*k*k*n => m*n). If the two inner k's are not the same number it can't cancel out and you will have an error.

[dav]

I checked it but I couldnt find where I was doing it wrong!

[dav]

The only place I could think of changing was

theta0 = [1; 1; 1; 1; 1; 1; 1];

but it didnt help