For loop and matrix indexing in matlab ques [closed]

Question

Closed. This question needs details or clarity. It is not currently accepting answers.

---

Want to improve this question? Add details and clarify the problem by editing this post.

Closed 8 years ago.

Improve this question

I have defined a matrix (initial_matrix) equal to 0 by using a for loop as:

I = 5;   % e.g number of nodes
for i =1:I
    initial_matrix = [0];    // an initial matrix will be generated for each node
end

Now, for each node i, I will consider all other nodes but not the node i itself and subtract each of them from 1 and take their product:

for node 1:

result = (1 - initial_matrix of node 2) * (1 - initial_matrix of node 3) * ...
    (1 - initial_matrix of node 4) * (1 - initial_matrix of node 5)

for node 2:

result = (1 - initial_matrix of node 1) * (1 - initial_matrix of node 3) * ...
    (1 - initial_matrix of node 4) * (1 - initial_matrix of node 5)

for node 3:

result = (1 - initial_matrix of node 1) * (1 - initial_matrix of node 2) * ...
    (1 - initial_matrix of node 4) * (1 - initial_matrix of node 5) 

and so..for the remaining 2 nodes!

Can any one tell me or give me hints on how this can be achieved? Thanks!

Answer 1

For each product (per node) you need to have all initial matrices in advance, so you should modify your initial loop to something along these lines:

initial_matrix = cell(I, 1);
for i = 1:I
    initial_matrix{i} = blah blah... %// Generated for each node
end

Then you can add another nested loop that does something like the following:

result = cell(I, 1);
for k = 1:I

    %// Compute product
    result{k} = 1;
    for n = setdiff(1:I, k)
        result{k} = result{k} * (1 - initial_matrix{n});
    end
end

Answer 2

I think these are the steps you need to take:

Assuming all matrices are of the same size and that you want to do elementwise multiplication:

1. Concatenate all of them into one big matrix initialMatrices, as desribed here.
2. Apply cumprod on 1-initialMatrices
3. If you used cumprod Extract the matrices that you need from this result

Some things to pay attention to:

• Maybe preprocess the first initial matrix by doing one minus it, or remove it.
• Make sure you build the total matrix in the right 'direction'. Check help cumprod for what the right direction is.

Answer 3

Currently, you are not creating an initial matrix for each node. If you want to have a separate one for each, I would suggest you to define a structure. So you will have:

I = 5;
for i =1:I
    initial_matrix.(sprintf('Node%d',i)) = [0];    % an initial matrix will be generated for each node
end

Then you can maybe do something like the following to do your operation (if I understood it right):

    for i =1:I
        numnode=[1:I]
        numnode(numnode==i)=[]
        for ind = 1:numel(numnode)-1
            NewMatrix.(sprintf('Node%d',i)) = (1- initial_matrix.sprintf('Node%d',numnode(ind))))*(1- initial_matrix.(sprintf('Node%d',numnode(ind+1))))
            end
    end

You can verify if this is correct by using a rand(1) statement instead of [0] while defining the initial_matrix.

I hope I understood you correctly