calculate mean af cell

I have a cell with 30 rows., In each row there are 39 columns. Every ten folders belong to one person. For each column I want to calculate the average value for the ten folders and give this value instead of 10 different amounts. then do it for folders from 11 to 20 and 21 to 30.

[Azzi Abdelmalek]

What do you mean by every ten folders?

[fereshte]

every 10 folders are including of 39 features.i want get mean of each features for 10 folders and put it instead relevant feature.

[Azzi Abdelmalek]

What do you mean by folder?

[Image Analyst]

By "folder" do you mean "row"????

[fereshte]

sorry not folder.cell mat contain 30 rows.and each row have 39 features. each 10 row Owned by an individual.but attribute values are different in 10 rowa for each person.for this reason,i want to find the mean values in each 10 rows and put them instead of the original values.

[Azzi Abdelmalek]

Each 10 rows from what?

[fereshte]

rows from 1 to 10 is person one.from 11 to 20 is person two.and 21 to 30 is person three.i want use this cell for person identification

Answer 1

[Azzi Abdelmalek]

clear out
m1=reshape(1:30,10,3);
for k=1:3
m=data_matrix(m1(:,k));
f=num2cell(reshape([m{:}],39,[])',1);
out{1,k}=arrayfun(@(x) mean(cell2mat(f{x}'),2),1:39,'un',0);
end
out=repmat(out,10,1)
out=out(:)

[fereshte]

i write this: load('data_all.mat')

out=cellfun(@(x) mean(reshape(x,10,3)),data_matrix,'un',0)

but get error: To RESHAPE the number of elements must not change.

Error in ==> @(x)mean(reshape(x,10,3))

Error in ==> Untitled3 at 2 out=cellfun(@(x) mean(reshape(x,10,3)),data_matrix,'un',0)

[Azzi Abdelmalek]

Look at edited answer

[fereshte]

in out must be 30 rows not 3 rows!!each 10 rows is similar togeather.

[Azzi Abdelmalek]

Then what mean do you want?

[fereshte]

each new 10 rows are mean of inital ten rows.

[Azzi Abdelmalek]

This is not clear

[fereshte]

i want do train and test with this data.in cell.mat there are 30 rows that each 10 rows Are related to one person but with diffrenet values.For every ten lines must be do mean that you do it perfect. but this mean value should be replaced in 10 rows instead of previous numbers.for 11 to 20 and 21 ta 30 too.output must have 30 rows. that each 10 rows are similar and contain mean valuse

[Azzi Abdelmalek]

Look at edited answer

[fereshte]

thank you so much.

Answer 2

[Roger Stafford]

If your "cell" is an ordinary 'double' 30 x 39 array, A, do this:

B = [mean(A(1:10,:),1);mean(A(11:20,:),1);mean(A(21:30,:),1)];

or else this:

B = reshape(mean(reshape(A,10,[]),1),3,[]);

Either method should work.

[fereshte]

i dont write this code.please help me :( my cell attached above its 30*1 cell.then in each cell there are 39 features that they are numerical. name of cell is data_all.mat. what write instead of A?