is it possible to change inner loops to make the script run faster?

hello

A small script I written takes a line and makes a cross correlation with all the lines after.

But it runs too slowly. could someone please advice?

Thanks.

clear Signal CROSS_FINAL MEAN_VAL stand_DIV
k=1;
num_run=1000;
start=440;final=1220; %indexes of the signal
for i=1:size(Untitled,2)
    Signal(:,i)=Untitled(start:final,i)-mean(Untitled(start:final,i));
end
for i=1:num_run
    i
    for j=i+1:num_run
        CROSS_FINAL(k,:)=xcorr(Signal(:,i),Signal(:,j));
        k=k+1;
    end
end
for i=1:final-start+1
    MEAN_VAL(i)=mean(CROSS_FINAL(:,i));
    stand_DIV(i)=std(CROSS_FINAL(:,i));
end

[Murugan C]

How does Untitled look , Is it Matrix data?

Can you please post Untitled data?

[Eli_Cog]

It looks like the file attached but with 1000 columns

Answer 1

[Guillaume]

None of the loops are needed:

for i=1:size(Untitled,2)
    Signal(:,i)=Untitled(start:final,i)-mean(Untitled(start:final,i));
end

is simply:

Signal = Untitled(start:final, :) - mean(Untitled(start:final, :), 1);

-

for i=1:num_run
    i
    for j=i+1:num_run
        CROSS_FINAL(k,:)=xcorr(Signal(:,i),Signal(:,j));
        k=k+1;
    end
end

is exactly what xcorr does when passed a matrix (I'm assuming that num_run is equal to size(Signal, 2), so:

CROSS_FINAL = xcor(Signal);

-

for i=1:final-start+1
    MEAN_VAL(i)=mean(CROSS_FINAL(:,i));
    stand_DIV(i)=std(CROSS_FINAL(:,i));
end

is simply:

MEAN_VAL = mean(CROSS_FINAL, 1);  %take mean across rows
stand_DIV = std(CROSS_FINAL, 1);  %take standard deviation across rows

So, overall, the whole code can be simplified to:

start=440;final=1220; %indexes of the signal
Signal = Untitled(start:final, :) - mean(Untitled(start:final, :), 1);   %take signal between start and final row and subract mean of each column
CROSS_FINAL = xcor(Signal);       %cross correlation of each column against each column
MEAN_VAL = mean(CROSS_FINAL, 1);  %take mean across rows
stand_DIV = std(CROSS_FINAL, 1);  %take standard deviation across rows

Note that this also removes the need for any clear.