How to segregate data according to its temperature value?

Hi all.

Suppose I have two 3d arrays, A and B, and both of them are of the same size (10000x100x100).

A is temperature values and B is water levels.

Values in A is associated with values in B of the same index. For example, the corresponding value for A(1,1,1) is B(1,1,1), the same goes for A(1,1,2) and B(1,1,2) and so on.

My aim is to segregate all the values in B into their respective temperature bins base on their temperature values in A.

All temperature bins are of one °C width. (E.g. any temperature value between 14 and 15 °C belongs to one group, and any temperature value between 15 and 16 °C belongs to another group.)

May I know how to do the above in an efficient manner? If all the above is too confusing, you can take a look at my code below. However, I feel that my code is not very efficient as there are a lot of 1d arrays and I am looking for ways to improve it. Thanks in advance.

My code:

max(A,[],'all'); %To find out the largest temperature bin value

min(nonzeros(A,[],'all'); %To find out the smallest temperature bin value

Map=cat(4,A,B);

size(Map);

%All the following variables are temperature bins

B_temp_21=[];

B_temp_22=[];

B_temp_23=[];

B_temp_24=[];

B_temp_25=[];

B_temp_26=[];

B_temp_27=[];

B_temp_28=[];

B_temp_29=[];

B_temp_30=[];

B_temp_31=[];

for i = 1:10000

for j = 1:100

for k = 1:100

if Map(i,j,k,1)>0

if fix(Map(i,j,k,1)) ==21

B_temp_21(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 22

B_temp_22(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 23

B_temp_23(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 24

B_temp_24(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 25

B_temp_25(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 26

B_temp_26(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 27

B_temp_27(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 28

B_temp_28(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 29

B_temp_29(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 30

B_temp_30(end+1)=Map(i,j,k,2);

elseif fix(Map(i,j,k,1)) == 31

B_temp_31(end+1)=Map(i,j,k,2);

else

continue

end

else

continue

end

end

end

i %Keep track of progress

end

[Torsten]

My aim is to segregate all the values in B into their respective temperature bins base on their temperature values in A.

What exactly do you mean by this ? For each temperature interval, you want to get the vector of water levels that belong to this temperature interval ?

[Baiwen Ding]

Yes this is what i meant sorry if it was confusing as i couldnt find any better words for it

Answer 1

[Torsten]

Use MATLAB's "discretize" for the temperature array and use the so received vector Y of bins to split the vector of water levels analogously:

https://de.mathworks.com/help/matlab/ref/double.discretize.html

Example:

rng('default')
A = rand(10,23,5);
B = rand(10,23,5);

X = linspace(0,1,10);
Y = discretize(A(:),X);
Bvec = B(:);
C = arrayfun(@(i)Bvec(Y==i),1:numel(X),'UniformOutput',0)

C = 1×10 cell array
    {132×1 double}    {139×1 double}    {121×1 double}    {131×1 double}    {127×1 double}    {130×1 double}    {140×1 double}    {110×1 double}    {120×1 double}    {0×1 double}

[dpb]

The only Q? then is do you want bins centered on even degrees or centered at mean between the even degree edges?

[Baiwen Ding]

Hi what do you mean by the only Q? Integers are the edges for the bins. Hope this clarifies.

[Baiwen Ding]

Hi i have tried the discretize function previously. However the code is still very iterative and long just like my current code. I am looking forwards to short my current code if possible. Thanks

[Torsten]

"discretize" is fast as fast can be.

So your complete loop and if constructions can be replaced by one call to "discretize", I guess.

Translated to your code:

X = floor(min(A(:))):ceil(max(A(:)));
Y = discretize(A(:),X);
Bvec = B(:);
B_temp = arrayfun(@(i)Bvec(Y==i),1:numel(X),'UniformOutput',0)

should substitute you complete code from above.