How to find the mean of a matrice based on a value in another matrice

Hello everyone,

I have two matrices (attached file data_x.mat contains two matrices/parameters) namely 'month' and 'sa'. I want to get, the mean of 'sa' which correspond to values, for example of month = 9 only. Anyone knows how to handle this?

Thanks

Answer 1

[Star Strider]

This returns all of the monthly means for all the months —

LD = load('data_x.mat')

LD = struct with fields:
    month: [125818x1 double]
       sa: [125818x1 double]

T1 = table(LD.month, LD.sa, 'VariableNames',{'month','sa'})

T1 = 125818x2 table
    month      sa  
    _____    ______

      9      34.532
      9       34.54
      9      34.545
      9      34.544
      9      34.542
      9      34.543
      9      34.544
      9      34.555
      9      34.542
      9      34.541
      9      34.541
      9       34.54
      9      34.532
      9      34.519
      9      34.527
      9      34.544

[Um,~,ix] = unique(T1.month, 'stable');
mmeans =  accumarray(ix, (1:numel(ix)).', [], @(x)mean(T1.sa(x)));

Monthly_Means = table(Um, mmeans, 'VariableNames',{'month','mean'})

Monthly_Means = 12x2 table
    month     mean 
    _____    ______

      9      34.551
     10      34.551
     11      34.552
     12      34.555
      1      34.557
      2      34.556
      3      34.555
      4      34.556
      5      34.555
      6      34.554
      7      34.552
      8      34.553

.

[Adi Purwandana]

Perfect! Thank you! One more question if you don't mind. Could you please assist me to make box plot for this results, so that I can see what's the stdev, mean, median, etc. to investigate how large the variance of the value for each month?

[Star Strider]

Thank you!

I am not certain how much information you can get from the boxchart, so I added a new accumarray call to produce that information. If you have the Statistics and Machine Learning Toolbox, you can also experiment with the boxplot function. I use boxchart here (introduced in R2020a) because you should have it as part of core MATLAB.

Try this —

LD = load('data_x.mat')

LD = struct with fields:
    month: [125818x1 double]
       sa: [125818x1 double]

T1 = table(LD.month, LD.sa, 'VariableNames',{'month','sa'})

T1 = 125818x2 table
    month      sa  
    _____    ______

      9      34.532
      9       34.54
      9      34.545
      9      34.544
      9      34.542
      9      34.543
      9      34.544
      9      34.555
      9      34.542
      9      34.541
      9      34.541
      9       34.54
      9      34.532
      9      34.519
      9      34.527
      9      34.544

[Um,~,ix] = unique(T1.month, 'stable');
monthdata =  accumarray(ix, (1:numel(ix)).', [], @(x){[mean(T1.sa(x)), median(T1.sa(x)),  std(T1.sa(x)), min(T1.sa(x)), max(T1.sa(x)), max(T1.sa(x))-min(T1.sa(x)), numel(T1.sa(x))]});

Results = array2table(cell2mat(monthdata), 'VariableNames',{'Mean','Median','StDev','Minimum','Maximum','Range','N'}, 'RowNames',compose('%3d',Um))

Results = 12x7 table
           Mean     Median     StDev      Minimum    Maximum    Range       N  
          ______    ______    ________    _______    _______    ______    _____

    9     34.551    34.553    0.015652    34.397     34.687     0.2908     9966
    10    34.551    34.553     0.01478    34.388     34.663     0.2757    13392
    11    34.552    34.553    0.012936    34.438     34.685     0.2475    12960
    12    34.555    34.555    0.011521    34.434     34.697      0.263    13392
    1     34.557    34.558    0.013831    34.367     34.733      0.367    13392
    2     34.556    34.557    0.012418    34.469     34.666     0.1963     9724
    3     34.555    34.555    0.014888    34.374     34.718     0.3433     8928
    4     34.556    34.557    0.012806    34.443     34.669     0.2257     8640
    5     34.555    34.556    0.013553    34.414     34.669     0.2552     8928
    6     34.554    34.553    0.012718    34.452     34.693     0.2403     8640
    7     34.552    34.551     0.01416    34.349      34.71     0.3613     8928
    8     34.553    34.554    0.014175    34.345      34.69     0.3441     8928

% Monthly_Means = table(Um, mmeans, 'VariableNames',{'month','mean'})

figure
boxchart(T1.month, T1.sa, MarkerStyle='+', JitterOutliers='on', Notch='on')
grid
xticks(1:12)
xlabel('Months')
ylabel('Data')
title('Box Plot')

Jittering the outliers produces a ‘violin plot’ effect, and gives a better sense of their distribution.

[Adi Purwandana]

Perfect as always! Thank you!

[Star Strider]

As always, my pleasure!

Answer 2

[Sameer]

Hi Adi

To calculate the mean of 'sa' values corresponding to a specific 'month' (e.g., month = 9), you can follow the below approach:

% Load the data from the .mat file
data = load('data_x.mat');

% Extract the 'month' and 'sa' matrices
month = data.month;
sa = data.sa;

% Find indices where month is equal to 9
indices = find(month == 9);

% Extract the corresponding 'sa' values
sa_selected = sa(indices);

% Calculate the mean
mean_sa = mean(sa_selected);

fprintf('The mean of sa values for month = 9 is: %.2f\n', mean_sa);

I hope this helps!

[Adi Purwandana]

Nice. Thank you @Sameer for your promptly help!