try to find max and min of x and y for each cluster

Hi, all

In my case, I have a table which stores x,y,and cluster numer in three columns. example as below

===================================================
608654.966062901  4820462.57604139  1
608662.024953254  4820455.91371599  1
608676.221786682  4820442.70145371  1
608676.221786682  4820442.70145371  1
608676.221786682  4820442.70145371  1
608676.221786682  4820442.70145371  1
608676.221786682  4820442.70145371  2
608676.221786682  4820442.70145371  2
608676.221786682  4820442.70145371  2
608676.221786682  4820442.70145371  2
608676.221786682  4820442.70145371  2
608676.221786682  4820442.70145371  2
608676.221786682  4820442.70145371  3
608676.221786682  4820442.70145371  3
608724.787042597  4820393.49671617  3
608732.086699384  4820386.94944425  3
608739.632386523  4820380.18398164  3
608747.011082827  4820373.74909310  3
608754.551408620  4820367.31682751  3
=====================================================
I am trying to find min/max of x and min/max of y in each cluster, and save them into a new table 'statcluster', as below
===============================================================
minX maxX minY maxY clusternum
===============================================================
************
Here is code I wrote, but it cannot work properly. some answers correct, but others are wrong. Please comment on it. Thanks alot
************
% find number of cluster -- cn
cn=max(dataa(:,3));
row=length(dataa(:,1));
% statcluster save max/min of x and max/min of y and their corresponding
% cluster number
statcluster=zeros(cn,5);
% compute number of elements in each cluster
trash=histc(dataa(:,3),unique(dataa(:,3)));

for i=1:row
    for j=1:cn
       if dataa(i,3)==j
%          min/max of x
           statcluster(j,1)=min(dataa(i,1));
           statcluster(j,2)=max(dataa(i,1));
%          min/max of y
           statcluster(j,3)=min(dataa(i,2));
           statcluster(j,4)=max(dataa(i,2));
%          cluster number
           statcluster(j,5)=j;
        break
       end
    end 
end

Answer 1

[Matt Fig]

One way to do this would be to use the ACCUMARRAY function.

DAT = [608654.966062901  4820462.57604139  1
       608662.024953254  4820455.91371599  1
       608676.221786682  4820442.70145371  1
       608676.221786682  4820442.70145371  1
       608676.221786682  4820442.70145371  1
       608676.221786682  4820442.70145371  1
       608676.221786682  4820442.70145371  2
       608676.221786682  4820442.70145371  2
       608676.221786682  4820442.70145371  2
       608676.221786682  4820442.70145371  2
       608676.221786682  4820442.70145371  2
       608676.221786682  4820442.70145371  2
       608676.221786682  4820442.70145371  3
       608676.221786682  4820442.70145371  3
       608724.787042597  4820393.49671617  3
       608732.086699384  4820386.94944425  3
       608739.632386523  4820380.18398164  3
       608747.011082827  4820373.74909310  3
       608754.551408620  4820367.31682751  3]

MX_x = accumarray(DAT(:,3),DAT(:,1) ,[],  @max);
MN_x = accumarray(DAT(:,3),DAT(:,1) ,[],  @min);
MX_y = accumarray(DAT(:,3),DAT(:,2) ,[],  @max);
MN_y = accumarray(DAT(:,3),DAT(:,2) ,[],  @min);

statcluster = [MN_x MX_x MN_y MX_y (1:3).']

[zheng]

everything is great, matt.

except the horizoncat happend to statcluster.

it said

Error using ==> horzcat

CAT arguments dimensions are not consistent.

[zheng]

btw, is it possible you can point out where I made wrong in my code?

Thanks alot

[Matt Fig]

Did you copy and paste my code exactly? Because I just did and got no horzcat error. I wonder if you are using data that is different than what I pasted above. Perhaps there are more than 3 groups in your data???

If not, try to copy EVERYTHING in the code I wrote above and paste, it should work. Then you need to figure out what is different about your data.

[zheng]

appreciate