for-loop to create vector

I am just learning for end loops.

I am trying to create a for end loops that can find the maximum value of each column of a matrix and store the results in a single vector.

I have ankle moment data and I want to identify the maximum moment of each stride (column in the data) using a for-loop to store the results in a vector.

Answer 1

[Ameer Hamza]

Since you insist on using for-loop

M; % your matrix
n = size(M, 2); % number of columns
max_vals = zeros(1, n);
for i = 1:n
    max_vals(i) = max(M(:,i));
end

Without for-loops

max_vals = max(M);

[Slane Haltine]

The code seems to work but I recieve an output of max_vals which is only one number, which is the maximum value. As opposed to a vector containing the max values of each column.

n = size(Force,10);
max_vals = zeros(1,n);
for i = 1:n
    max_vals(i) = max(Force(:,i));
end

[Rik]

Have you read the documentation for the size function? You're asking for the tenth dimensions. I doubt you have that many.

[Slane Haltine]

Okay, I was confused by the original code stating %number of columns

Thank you, I have read it but was misinterpretting it before.

In that case it absolutely works but how can I have the output as a column vector instead of a a row?

[Ameer Hamza]

If you want output to be in column vector form then initialize like this

max_vals = zeros(n,1);

[Slane Haltine]

Thank you very much!

[Ameer Hamza]

I am glad to be of help!

[Slane Haltine]

I am a bit confused about the use of zeros in this line of code. The code works perfectly, but can you elaborate as to what this line is calling for?

[Ameer Hamza]

Read about pre-allocation in MATLAB. It is just there to make the code more efficient. Removing it will not cause any error but the code can get very slow if the array needs to be large.

[Rik]

Have you read the documentation? It is one of the major benefits of Matlab over competing software.

[Slane Haltine]

I think I understand and I see how it can make the code much more efficient for a much larger array. If i remove it, the values are given in a row vector again.

This is what confuses me as to its function in this code.

for example, without it, this is how I would write the code.

n = size(Force,2); 
max_vals = 0;
for i = 1:n       
    max_vals(i) = max(Force(:,i));
end
max_column = reshape(max_vals,10,1)

[Rik]

You can make it a row vector if you prefer. It is good to form the habit when you learn, so you don't need to unlearn bad habits later.

[Slane Haltine]

I mean to say, I added the reshape line at the end of the code because removing the zeros line results in a row instead of a column. That confuses me as to what the zeros command is doing in that line of code and how it makes the difference between the max_vals resulting in a row or column vector.

[Rik]

If you don't specify two coordinates, Matlab will attempt to figure out which dimensions should be extended to make your data fit. The default is to grow in the first dimension. If you create the array with a specific shape, Matlab will extend that. So you can choose between zeros(1,n); or zeros(n,1); (or more exotic shapes like zeros(1,1,n);).

[Slane Haltine]

Awesome, thank you!

[Ameer Hamza]

Note that If you just want to conver a row vector to a column vector then you can even use

max_column = max_column(:);
% or
max_column = max_column.'; % in this case, max_column must be a row vector