How to average 5 rows of a matrix iteratively?

1 view (last 30 days)
Samuel White
Samuel White on 5 Sep 2018
Commented: Star Strider on 5 Sep 2018
I have a large matrix. I wish to avaerage the data from 5 rows and write the values to a new matrix iteratively until the whole matrix has been averaged. How can I do this? Many thanks.
  4 Comments
Show 2 older comments
Star Strider
Star Strider on 5 Sep 2018
The matrix you posted has 44 rows and 10 columns.
What result do you want?
Samuel White
Samuel White on 5 Sep 2018
Well, it actually has 5000 rows and 10 collumns. I want to average sets of 5 rows ideally :)

Sign in to comment.

Sign in to answer this question.

Answers (2)

Joel Miller
Joel Miller on 5 Sep 2018
I am unsure whether you want to average rows 1-5, 2-6, 3-7, etc., or rows 1-5, 6-10, 11-15, etc. I assume the number of rows in your array, y, is an integer multiple of 5. You can create a 3D array to get a solution for the former case:
for k=1:5
z(:,:,k)=circshift(y,1-k,1);
end
z=z(1:end-4,:,:);
m=mean(z,3);
Likewise, for the latter case, create a 3D array:
for k=1:size(y,2)/5
z(k,:,:)=permute(y((k-1)*5+1:(k-1)*5+5,:),[3 2 1]);
end
m=mean(z,3);
Star Strider
Star Strider on 5 Sep 2018
This creates ‘blocks’ of 5 rows each, then takes the mean down the columns (across rows) of each ‘block’. I shortened the matrix you posted by 4 rows (to make the number of rows an integer multiple of 5), so I am re-posting it here as well as the code.
M = [0 0.1904 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1904 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1904 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0.0002 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0100 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0098 0 6.5535
0 0.1903 -0.0235 0.0690 0.0106 -0.0001 -0.0179 0.0104 0 6.5535
0 0.1902 -0.0235 0.0691 0.0106 -0.0001 -0.0179 0.0095 0 6.5535
0 0.1901 -0.0235 0.0691 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1900 -0.0236 0.0692 0.0106 -0.0001 -0.0179 0.0099 0 6.5535
0 0.1899 -0.0236 0.0692 0.0106 -0.0001 -0.0179 0.0099 0 6.5535];
M3 = reshape(M', size(M,2), 5, []); % Reshape ‘M’
M3 = permute(M3, [2 1 3]); % Change Dimension Order
Mmean5 = squeeze(mean(M3,3)) % Take Row Means Of 5-Row ‘Blocks’
Experiment to get the result you want.
  2 Comments
Samuel White
Samuel White on 5 Sep 2018
Thank you for your response. Unfortunately I am quite new to matlab and so I'm not sure I quite understand. I tried the code but it keeps returning an error:
Error using reshape Product of known dimensions, 45, not divisible into total number of elements, 24957.
Error in filefix (line 3) M3 = reshape(M', size(M,2), 5, []); % Reshape M
Many thanks
Star Strider
Star Strider on 5 Sep 2018
My pleasure.
My code returns the mean values of ‘blocks’ of 5 rows, and I shortened your matrix to 40 rows (an integer multiple of 5) so that it would be compatible with 5-row blocks.
The matrix you are working with has 2 incompatible rows:
ExtraRows = rem(24957, 5)
ExtraRows =
2
Probably the easiest way to deal with that is:
M3 = reshape(M(1:24955,:)', size(M,2), 5, []); % Reshape ‘M’
M3 = permute(M3, [2 1 3]); % Change Dimension Order
Mmean5 = squeeze(mean(M3,3)) % Take Row Means Of 5-Row ‘Blocks’
then:
Last2 = mean(M(end-1:end,:))
You can then append (concatenate) ‘Last2’ to the end of ‘Mmean5’ as:
Mmean5 = [Mmean5; Last2];
or just keep them separate.
This last is UNTESTED CODE, since I do not have your matrix. It should work.

Sign in to comment.

Categories

Find more on Numeric Types in Help Center and File Exchange

Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!