How to take the maximum value of an array?

Hi, I have this array a = [1:20]. How do I take the maximum value of every 5 elements of it. So I will have b = [5 10 15 20]. Anyone can help me? Thank you.

Answer 1

[Star Strider]

This works:

a = [1:20];
b = max(reshape(a, 5, []));

[Greig]

And will be much faster and less typing than my solution! Kudos!

[arina octave]

hi, thank you for kindly answer my question. it is really effective. i've tried your code like this :

but what i want is the x axis of signal b exactly like the x axis of signal a. do you have any idea how to do that? thank you...

[Star Strider]

My pleasure.

Yes, I do actually have an idea of how to do it, because I wrote the code for your earlier answer. See if this does what you want:

N = 250;                                    % Length Of Vector
t = linspace(0, 20, N);                     % Time
sig = rand(1,N);                            % Original Signal
mtx = max(reshape(sig, 5, []));
newsig = bsxfun(@times, mtx, ones(5, size(mtx,2)));
newsig = reshape(newsig, [], length(sig));  % New Signal

figure(1)
subplot(2,1,1)
plot(t, sig)
grid
subplot(2,1,2)
plot(t, newsig)
grid

It works with my test vector, see if it works with your signal.

[Star Strider]

@Greig — It’s a code shortcut of the sort that frequently appears here on ‘Answers’. I’m frequently amazed at the Answers the truly gifted people here submit.

[arina octave]

and I'm truly amazed with your answer.. aaa thank you so much Star Strider!

[Star Strider]

My pleasure!

Answer 2

[Greig]

Try something like this...

a=1:20;
steps = 5;
b = NaN(mod(length(a)/steps,steps), 1); % preallocate b for speed, if you want to run larger loops
count = 1; % an index count for b in the loop

for ii = 1:steps:length(a)-(steps-1)
    inds = ii:ii+(steps-1); % The indices we are interested in
    b(count) = max(a(inds)); % Get the max
    count = count+1; % add 1 to the counter
end

[arina octave]

hi, thank you for kindly answer my question. but what i want is the x axis of signal b exactly like the x axis of signal a. do you have any idea how to do that? thank you...