How to calculate the mean of an interval (from a vector) in a loop?

Hey,

I would like to calculate the mean in intervals of the vector total_sync_spk, but i dont know how to do it. My idea when doing it manually looks like this:

mean(total_sync_spk(1:10))

mean(total_sync_spk(11:20)) and so on ...

This is my script:

all_means = [];

all_stds = [];

total_sync_spk = [];

for Gsyn = 0:2:20

for i = 1:10

[spt1, spt2] = IFcoupling(300, 300, Vth1, Vth2, ...

Tref1, Tref2, Esyn1, Esyn2, Gsyn, Gsyn, stim_time, dt);

[Nco, xidx, yidx] = coincidenceCounterSophisticated(spt1,spt2,0.25);

total_sync_spk = [total_sync_spk Nco];

end

a_means = mean(total_sync_spk);

all_means = [all_means a_means];

a_stds = std(total_sync_spk);

all_stds = [all_stds a_stds];

end

[Mathieu NOE]

hello Paul

first, you could probably index the all_means and all_stds (vectors) instead of doing concatenation, I mean in the outer loop :

all_means(Gsyn) = mean(total_sync_spk);

instead of the 2 lines

    a_means   = mean(total_sync_spk);
    all_means = [all_means a_means];

then you can define an output for each individual segment like :

    all_means1(Gsyn) = mean(total_sync_spk(1:10));
    all_means2(Gsyn) = mean(total_sync_spk(11:20)); 

but this could be expanded in more generic way using another for loop (how many indexes ?)

Answer 1

[Shashank Gupta]

Hi Paul,

There is one more convenient way of doing this, you can reshape your array in a matrix of shape (rows,10) such that each column i contains total_sync_spk(i:i+10) and then call the mean function column wise and it will do your job. I can attach a small peice of code for your reference.

% Generate an array. 
arr = 1:100';

% Reshape the array in (some_rows,10) format.
arr_reshape = reshape(arr,10,10);

% Take the mean column wise.
mean_arr_reshape = mean(arr_reshape,1);

I hope this helps.

Cheers.