Grouping or categorizing data based on conditions

Question

Hamid 2022-7-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1768635-grouping-or-categorizing-data-based-on-conditions

回答： Hamid 2022-8-10

采纳的回答： Hamid

在 MATLAB Online 中打开

Dear all,

In my calculations data is generated as attached which is x,y,z and a value as 4th dimension.

I wanted to divide this data into goups based on their value for x,y and z and assign one index to each category and finally find how many points I have in each group. It's a kind of dividing a region into voxels for 3d or pixels in 2d.

For example, if 20 < x < 19 & 20 < y< 19 & 20 <z<19 => set 1 index for the data which are located in this category.

I found these functions but don't know which one is approporiate for my purpose as the data is huge; findgroups,categories(A) ,[X,Y,Z] = meshgrid(x,y,z), Y = discretize(X,edges).

Thank you in advance.

data = readtable ('Outpu.txt');

data = table2array(data);

x=data(:,1);

y=data(:,2);

z=data(:,3);

t=data(:,4);

figure

scatter3(x,y,z,2,t,'filled')

ax = gca;

ax.XDir = 'reverse';

zlim([-200 200])

xlim([-100 100])

ylim([-100 100])

view(-31,14)

xlabel('X')

ylabel('Y')

zlabel('Z')

cb = colorbar;

clim([0 20])

grid minor

3 个评论
显示 1更早的评论隐藏 1更早的评论

Hamid 2022-7-27

@KSSV Thank you so much for your reply.

As you see in the plot, I wanted to divide all the region showed in the graph based on x and y and z into sub-regions (voxels) and set one index to each. Finally, calculate the number of points and average of t in each voxel.

I hope the following figure make my question clear.

Chunru 2022-7-27

See this: https://www.mathworks.com/matlabcentral/fileexchange/29435-n-dimensional-histogram-count

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Hamid 2022-8-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1768635-grouping-or-categorizing-data-based-on-conditions#answer_1024185

在 MATLAB Online 中打开

for i = 1:length(xedges)-1
    for j = 1:length(yedges)-1
        for k = 1:length(zedges)-1
                idx = ((xedges(i)<= x_p) & (x_p < xedges(i+1))) & ((yedges(j)<= y_p) & (y_p < yedges(j+1))) & ((zedges(k)<= z_p) & (z_p < zedges(k+1))) ;
                G(i,j,k) = nnz(idx);                          
        end
    end
    
end

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 2

KSSV 2022-7-27

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1768635-grouping-or-categorizing-data-based-on-conditions#answer_1015835

在 MATLAB Online 中打开

How about this?

data = readtable ('https://in.mathworks.com/matlabcentral/answers/uploaded_files/1078835/Outpu.txt');

data = table2array(data);

x=data(:,1);

y=data(:,2);

z=data(:,3);

t=data(:,4);

x0 = linspace(-40,40,5) ;

y0 = linspace(-40,40,5) ;

z0 = linspace(-200,200,10) ;

G = NaN(length(x),1) ;

count = 0;

for i = 1:length(x0)-1

for j = 1:length(y0)-1

for k = 1:length(z0)-1

count = count+1 ;

idx = ((x0(i)<= x) & (x < x0(i+1))) & ((y0(j)<= y) & (y < y0(j+1))) & ((z0(k)<= z) & (z < z0(k+1))) ;

G(idx) = count ;

end

%% Remove the points

x(isnan(G)) = [];

y(isnan(G)) = [];

z(isnan(G)) = [];

G(isnan(G)) = [] ;

figure

scatter3(x,y,z,2,G,'filled')

colormap(turbo)

5 个评论
显示 3更早的评论隐藏 3更早的评论

KSSV 2022-7-28

You can get the count from G also. nnz(G==1) gives the number of points in the first cube etc.

Hamid 2022-7-29

在 MATLAB Online 中打开

It returns 0 again. Could you please take a look at the code again? Thank you in advance.

data = readtable ('https://in.mathworks.com/matlabcentral/answers/uploaded_files/1078835/Outpu.txt');
data = table2array(data);
x=data(:,1);
y=data(:,2);
z=data(:,3);
t=data(:,4);
x0 = linspace(-40,40,5) ;
y0 = linspace(-40,40,5) ;
z0 = linspace(-200,200,10) ;
G = NaN(length(x),1) ;
count = 0; 
for i = 1:length(x0)-1
    for j = 1:length(y0)-1
        for k = 1:length(z0)-1
            count = count+1 ; 
            idx = ((x0(i)<= x) & (x < x0(i+1))) & ((y0(j)<= y) & (y < y0(j+1))) & ((z0(k)<= z) & (z < z0(k+1))) ;
            G(idx) = count ; 
            cnt = nnz(G==1) ; %%%%% I added this line but it returns zero.
        end
    end
end
%% Remove the points
x(isnan(G)) = [];
y(isnan(G)) = [];
z(isnan(G)) = [];
G(isnan(G)) = [] ;
figure
scatter3(x,y,z,2,G,'filled')    
colormap(turbo)