How to create a For Loop to Loop through all the rows in the matrix

Hello all. I am working with time series data. I have a matrix 142x240000 (142 rows, each row has 240000 columns). I want to calculate the the normalisation weight for each row. Below is the code:

%Normilisation time window width.half of the max period of the bandpass which is applied to the data prior to cross-correlation. (7Hz-40Hz)
N=0.07;
%Calculate the size of the signal to get number of columns and rows. n is number of rows and p is number of columns
[n,p] = size(signal);
%Calculate the normalised weights
count = 0
for j = 1:142 %Matrix indexing, each row representing 10 minutes
  count = count + 1;
  ten = 240000; %Number of columns
  %Divide signal into 142 rows and 240000 columns
  A(count,:) = signal(((j-1)*ten)+1:ten*j);

The above code runs just fine. It takes the signal that I have and divides it into 142 rows and 240000 columns. I now need to calculate the normalisation weight for each row. In the end, I want to have a matrix called weight with the size 142x1, each row should have the calculated weight of it's corresponding row from matrix A. The formula for calculating weight is:

weight =(1/((2*N)+1))*sum(abs(row)).

The code I have to calculate the weight is:

    for i = 1:142
      weight(i)=(1/((2*N)+1))*sum(abs(A)); %The A here should be the rows(row1,row2,...till row 142)
    end

How do I make the For Loop to calculate the weight from row 1 to row 142 and store the calculated weights as a 142x1 matrix

Answer 1

[dpb]

You don't need any loops at all...use the MATrix LABoratory vectorized operations. You already have the the signal in the array, you don't need anything to use it--

weight=1/(2*N+1)*sum(abs(signal),2);

The only "trick" here is to use the optional second argument to sum to add row elements (2nd dimension) instead of the default by column.

NB: The above assumes your parentheses-nesting is correct as the formula is written that the weight is intended as

wt = SUM(|signal|)/(2N+1)

and not as

wt = 1/{(2N+1)*SUM(|signal|)}

If it is instead the latter, you'll need

weight=1./((2*N+1)*sum(abs(signal),2));

using the "dot" divide operator for element-wise division.

[Tee-Kay]

Thank you.