Binning matrix data (negative and positive decimals): mean of each 100 rows and create a new matrix

Question

William Norbey Sanchez Luna 2024-8-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2147999-binning-matrix-data-negative-and-positive-decimals-mean-of-each-100-rows-and-create-a-new-matrix

评论： William Norbey Sanchez Luna 2024-8-29

I have a A= 7382300X12 matrix. I want to reduce it by averaging every 1000 rows in all columns. So at the end I will have a B= 73823X12 matrix.

The matrix A have negative and positive decimals, so blockproc() didn't work for me.

How can I do it?

2 个评论
显示无隐藏无

Walter Roberson 2024-8-26

You have the small issue that 7382300 is not evenly divisible by 1000. Should the final 300 entries be treated as a complete group?

Matt J 2024-8-26

编辑：Matt J 2024-8-26

The matrix A have negative and positive decimals, so blockproc() didn't work for me.

Whatever went wrong, that wasn't the reason.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Sahas 2024-8-27

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2147999-binning-matrix-data-negative-and-positive-decimals-mean-of-each-100-rows-and-create-a-new-matrix#answer_1505699

在 MATLAB Online 中打开

Hi @William Norbey Sanchez Luna,

As per my understanding, you would like to reduce the size of an input matrix by taking the mean of every 1000 rows and all the columns.

The following code snippet achieves the desired functionality. You can modify the input matrix to suit your preferred data types.

inputMatrix = randi([-100, 100], 3300, 12);
numRows = size(inputMatrix, 1);
numCols = size(inputMatrix, 2);
blockSize = 1000;
% Calculate the number of complete blocks
numBlocks = floor(numRows / blockSize);
% Reshape inputMatrix to group every 'blockSize' rows together for each column
reshapedA = reshape(inputMatrix(1:blockSize*numBlocks, :), blockSize, numBlocks, numCols);
% Calculate the mean along the first dimension
semiFinalMatrix = squeeze(mean(reshapedA, 1))
semiFinalMatrix = 3x12
   -0.1360    1.0220   -1.5510   -0.2930   -1.3370    0.6940    2.4430    0.1380   -5.4680   -2.2770   -2.1510    0.8750
   -0.2440    0.1370    1.8730   -1.8800   -1.3020    0.7470   -1.5090   -0.5750   -1.6570   -3.6780    1.0780   -0.3620
    0.7000   -0.9950    2.4580   -0.5420    2.3390    1.2080   -4.0240    0.7980   -0.9520    1.9370    2.7380    1.2630
<mw-icon class=""></mw-icon>
<mw-icon class=""></mw-icon>
% Assuming treating the leftover rows as one group and getting their mean as well
remainingRows = mod(numRows, blockSize);
if remainingRows > 0
    % Calculate the mean of the remaining rows
    remainingMean = mean(inputMatrix(end-remainingRows+1:end, :), 1);
    
    % Append the mean of the remaining rows to semiFinalMatrix
    finalMatrix = [semiFinalMatrix; remainingMean]
end
finalMatrix = 4x12
   -0.1360    1.0220   -1.5510   -0.2930   -1.3370    0.6940    2.4430    0.1380   -5.4680   -2.2770   -2.1510    0.8750
   -0.2440    0.1370    1.8730   -1.8800   -1.3020    0.7470   -1.5090   -0.5750   -1.6570   -3.6780    1.0780   -0.3620
    0.7000   -0.9950    2.4580   -0.5420    2.3390    1.2080   -4.0240    0.7980   -0.9520    1.9370    2.7380    1.2630
    2.6933   -2.9067    5.5567   -1.9933   -2.2433    3.5667   -1.0933   -6.3200    0.4633    4.0300    0.3400    0.2767
<mw-icon class=""></mw-icon>
<mw-icon class=""></mw-icon>