Reference to non-existent field

Hello i'm new to matlab, this is my error "Reference to non-existent field 'sigmaV'."

This is my code:

%Script tests the particle filter 

%simulation parameters
%[sigma sigmaN mu F/V_end]
param_real=[.25 0.0001 .18 1];  

sigma=param_real(1);   %asset volatility
delta=param_real(2);   %measurement noise
mu=param_real(3);       %asset drift

F0=1;                   %debt
V_end=F0/param_real(4); %end-of-sample pseudo-leverage

r=.05;                  %riskfree rate
dt=1/250;               %observation frequency

T=10; %beginning maturity
t=9; %end-of-sample maturity

%number of simulated paths
Npath=1;

randn('state',0);

%%%%%%%%%%%%%%%%
%Simulate paths%
%%%%%%%%%%%%%%%%

for m=1:Npath

  [S(:,m),AssetValue(:,m),mat]=SimMerton_Equity(V_end,dt,mu,r,sigma,F0,T,t,delta);

end

%%%%%%%%%%%%%%%%%%%%%%%
%set filter parameters%
%%%%%%%%%%%%%%%%%%%%%%%

samplesize=length(mat);

F=ones(samplesize,1)*F0;

modelparam.dt=dt;
modelparam.r=r;
modelparam.sigma=sigma/20;
modelparam.mu=mu;
modelparam.delta=delta;

%number of particles 
M=1000;
R=1000;

%random number seed to filter
seed=0;

%generate prior
Prior=log(AssetValue(1))+randn(M,1)*.1;

%%%%%%%%%%%%
%Run Filter%
%%%%%%%%%%%%

tic

[L,V]=feval('LocalizedFilter',modelparam,Prior,R,seed,S(1:100),F(1:100),mat(1:100));

toc

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%loglikelihood from transformed data method%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

L3=Merton_loglvec([modelparam.sigmaV modelparam.mu],S(1:100),r,F(1:100),mat(1:100),modelparam.dt);

Answer 1

[Image Analyst]

I see you assigned

modelparam.sigma=sigma/20;

but where exactly did you assign

modelparam.sigmaV = ?????????????????????

I'm not seeing it. You need to assign it some value. Or maybe you want to use modelparam.sigma instead of modelparam.sigmaV???

[Jony Bench]

This is the code from an author, so i really dont know what's the problem, i forgot to say i removed V from sigmaV and i got these errors:

Error using  + 
Matrix dimensions must agree.

Error in GetV (line 4)
V_high=S_observed+F;

Error in Merton_loglvec (line 8)
[V,d]=GetV(S_observed,r,sigmaV,F,mat);

[Jony Bench]

and these are the functions GetV and Merton_Loglvec

function [V,d]=GetV(S_observed,r,sigmaV,F,mat,MAXITER,tol)

V_low=S_observed;
V_high=S_observed+F;

%compute equity values from Merton's model

iter=0;
stopcrit=1;

if nargin<7
    tol=.00000001;
end

if nargin<6
    MAXITER=1000;
end

while (stopcrit>0 & iter<MAXITER)

    V_new=(V_low+V_high)/2;
        
    %evaluate S at V_new
    d=(log(V_new)-log(F)+(r+sigmaV^2/2).*mat)./(sigmaV*sqrt(mat));

    S=V_new.*normcdf(d)-F.*exp(-r.*mat).*normcdf(d-sigmaV*sqrt(mat));
       
    k1=find(S<S_observed);
    
    V_low(k1)=V_new(k1);
    
    k2=find(S>=S_observed);
    
    V_high(k2)=V_new(k2);
    
    stopcrit=max(abs(V_high-V_low))>tol;
    
    iter=iter+1;
    
end
    
if (iter==MAXITER)
    warning('MAXITER reached in GetV')
end

V=V_new;

---------------------

function [logl_vec,lnV]=Merton_loglvec(param,S_observed,r,F,mat,dt)
%PURPOSE:   returns vector of loglikelihood values for Merton's model using transformed data method


sigmaV=param(1);
mu=param(2);

[V,d]=GetV(S_observed,r,sigmaV,F,mat);

samplesize=length(S_observed);

lnV=log(V);


logl=-1/2*log(2*pi)-1/2*log(sigmaV^2*dt)...
    -(lnV(2:end)-lnV(1:end-1)-(mu-sigmaV^2/2)*dt).^2/(2*sigmaV^2*dt);

%get determinant of jacobian
detJ=-log(normcdf(d(2:end)))-lnV(2:end)+log(S_observed(2:end));

logl_vec=logl+detJ;

[Image Analyst]

Evidently S_observed and F don't have the same lengths. It says that. What does it say in the workspace that their lengths are?

Debugging in MATLAB | Doug's MATLAB Video Tutorials