Storing a Matrix Output in a matrix

%% I am running the following code but having some error:

prevalence=zeros(iters,1+max(data.t));

for i=1:iters

% copy existing parameter vector

proposal=theta;

% Generate a new proposal for epsilon

proposal=normrnd(theta(1),sigma);

% check proposal is in range

if proposal(1)>0

% calculate log acceptance ratio (may be bigger than 1)

lar = logLikelihoodSIR2(proposal,data,N) - logLikelihoodSIR2(theta(1),data,N);

% don't forget the prior ratio!

lar = lar + logPrior(proposal) - logPrior(theta(1));

% generate a random number between 0 and 1;

u = unifrnd(0,1);

% accept if lar>log(u) iff ap>u

if lar>log(u)

% the proposal becomes the new value of theta

theta=proposal;

accept=accept+1;

else

reject=reject+1;

end

else

% automatically reject outside the range for beta (as it has prior

% density zero).

reject=reject+1;

end

% store parameters for output every iteration (at the moment)

stored(i,:)=theta;

% recalculate and store trajectory for output

Numgroups = 5;

N = [600000 800000 1000000 1500000 6000000];

sigma = [0.01 0.05 0.2 0.3 0.4];

para = struct('N',N,'H',0.25,'da',0.3,'tau',0.15,'gamma',0.1, ...

'epislon',theta(1),'Numgroups',Numgroups,'sigma',sigma,'q_s',0.05, ...

'q_w',0.2, 'q_o',0.5,'q_h',1/4,'phi_t',0.25,'theta',0.45);

S = [600000 800000 1000000 1500000 6000000] - [1 zeros(1,Numgroups - 1)];

%Define initial conditions as a structure

ICs = struct('S',S,'E_f',[1 zeros(1,Numgroups - 1)],'E_sd',zeros(1,Numgroups), ...

'E_su',zeros(1,Numgroups),'E_q',zeros(1,Numgroups),'D_f',zeros(1,Numgroups), ...

'D_sd',zeros(1,Numgroups),'D_su',zeros(1,Numgroups),'D_qf',zeros(1,Numgroups), ...

'D_qs',zeros(1,Numgroups),'U_f',zeros(1,Numgroups),'U_s',zeros(1,Numgroups), ...

'U_q',zeros(1,Numgroups),'N',N);

Classes=ODE_agemodelkeelMCMC(para,ICs,min(data.t),max(data.t));

prevalence(i,:) = Classes.E_f./N;

end

My question is in the BOLDEN PART.

Classes.E_f./N is an (844 by 5) matrices which is what I want.

Prevalence(i,:) is a (1 by 844) which means that I want to store each iteration of Classes.E_f./N into the prevalence matrix I have pre-defined. But I am running into error : Unable to perform assignment because the indices on the left side are not compatible with the size of the right side.

I know the sizes differs but I don't know how better to handle this.

Answer 1

[Steven Lord]

You have 4220 (= 844*5) eggs and an egg carton with 844 cups. How can you fit all 4,220 eggs into 844 cups (and no, scrambling the eggs is not an option in this metaphor.)

Either you need to cut down the amount of data you're trying to store in Prevalence(i, :) or "you're gonna need a bigger matrix."

I'd probably preallocate Prevalence to be an array of size [844 5 niter] and store the matrix of parameters in each page of the array, assuming that the size of Classes.E_f./N is always [844 5].

A = [1 2; 3 4];
Z = zeros(2, 2, 3);
for k = 1:3
    fprintf("k is %d\n", k)
    Z(:, :, k) = A^k
end

k is 1

Z = 
Z(:,:,1) =

     1     2
     3     4


Z(:,:,2) =

     0     0
     0     0


Z(:,:,3) =

     0     0
     0     0

k is 2

Z = 
Z(:,:,1) =

     1     2
     3     4


Z(:,:,2) =

     7    10
    15    22


Z(:,:,3) =

     0     0
     0     0

k is 3

Z = 
Z(:,:,1) =

     1     2
     3     4


Z(:,:,2) =

     7    10
    15    22


Z(:,:,3) =

    37    54
    81   118

[Gbeminiyi Oyedele]

Thank you. this helps