Finding the Maximum of 6 matrices

I have 6 41*41 matrices. They contain peak displacement data for a combination of variables which I have changed

I.e. I would have a matrix which would look like this for example: where the kiso values are going down and the ciso vaules are going across and the values are the correspsonding data for each combination.

M1=
         ciso
  kiso     1     2    3    
  1        3.4   3    0.5  
  2        8     5.6  0.23 
  3        2     4.3  0.56  

M2=
         ciso
  kiso     1     2    3    
  1        4     5    6  
  2        0.1   5.2  0.2 
  3        9     4  0.5 

Max=

           ciso
    kiso     1     2    3    
    1        4     5    6 
    2        8     5.6  0.23 
    3        9     4.3  0.56

How do I compare each matrix and build a separate matrix which has the maximum value for each kiso and ciso combination as I have done in the example?

I have tried the matrix max comparison max(A,B) etc but I get an error saying there are too many input arguments as I am comparing 6 matrices not 2. max_building=max(Peak_mat1,Peak_mat2,Peak_mat3,Peak_mat4,Peak_mat5,Peak_mat6);

Thanks in advance

Answer 1

[njj1]

You can use max(X,[],dim) to specify the dimension from which to compute the maximum value. For example, if you let dim=2, then this computes the maximum value across each row of your matrix, resulting in a column vector; while dim=1 takes the maximum over each column of your matrix, resulting in a row vector. You could also try to re-arrange your data into a 41x41x6 array, where each "slice" in this 3-D cube is one of your original matrices. If you then entered max(X,[],3), you would get a 41x41 matrix where each entry is the maximum value of that particular (i,j) location for each of the 6 41x41 matrices. You could then pare this down into a maximum value for all columns and all rows over all the 6 matrices. One way to do this is: max(max(X,[],3),[],2) to get the max across each row for all 6 arrays. To do this same thing for columns: max(max(X,[],3),[],1). Hope this helps!

[serena solanki]

Thank you very much for this!

I will try and combine the matrices into a 41*41*6 array. I am a little unsure how to save the data into an array as I want to make sure its in the right order.

Thanks,

Serena

Answer 2

[dpb]

The problem is the creation of sequentially named variables instead of using a form of data storage that makes addressing the data easy programmatically.

The simplest conceptually for the 2D arrays would be to create a 3D array where each of the six is a plane (3rd dimension); then you can simply address each by the desired index 1 thru 6.

If you create the M array in that fashion for the above case you'd have

M(:,:,1) =
  3.4000    3.0000    0.5000
  8.0000    5.6000    0.2300
  2.0000    4.3000    0.5600
M(:,:,2) =
  4.0000    5.0000    6.0000
  0.1000    5.2000    0.2000
  9.0000    4.0000    0.5000
>>

and the desired max are simply

>> max(M,[],3)
ans =
  4.0000    5.0000    6.0000
  8.0000    5.6000    0.2300
  9.0000    4.3000    0.5600
>>

[serena solanki]

So as long as I call each matrix M(:,:1)----M(:,:,6) I can use the max(M,[],3) ?? or would I have to use 41 i.e. max(M,[],41) as I have 41 columns ??

I just want to make sure I have understood what you have said.

Thanks,

Serena

Answer 3

[Siyu Guo]

Guess max(cat(3,M1,M2,M3,M4,M5,M6), [], 3) should do.