Automatically adjust bin width

Dear all,

I would like to do the following:

1. Bin TM_3 using bin width of 2 and select the coincident R_3.
2. The binned R_3 should have >= 150 data points for each bin.
3. If < 150, the bin width is extended until 150 data points is obtained.

I am having problems with extending until 150 data points are obtained. I have attached my code below. If there are other better methods, please let me know. Thank you very much for your time.

clear; clc;
load('Test_CC.mat');

TP     = [TM_3 R_3];
TP_s   = sortrows(TP);                                                      % Sort ascending based on Tm
bin_a1 = (TP_s(1:end-2,1));                       
bin_b1 = (TP_s(1:end-2,1)+2);  

% % Initial binning according to specified width ==========================
for i = 1:length(TP_s)
    for j = 1:length(bin_b1)
        bin_c1 = find(TP_s(:,1)>= bin_a1(i) & TP_s(:,1) <= bin_b1(i));      % Initial bin of TM_3    
        bin_d1 = length(bin_c1);                                            % Length of initial bin
    if bin_d1 < 150                                                         % Check if binned data < 150
        bin_c2 = find(TP_s(:,1)>= bin_a1(i) & TP_s(:,1) <= bin_b1(j));      % Try different end bin until 150
        if length(bin_c2) == 150                                            % Check if new binned data >= 150
            bin_b1(i,:) = bin_b1(j);                                        % New end bin
            break;                                                          % If >=150, break.
        end
    end
    end
    bin_c3{i,:}   = bin_c2;
    bin_a1(i+1,:) = bin_b1(i)-1;
    bin_b1(i+1,:) = bin_a1(i+1)+2;
end
bin_a2 = bin_a1(find(bin_b1 < max(T_mx)));
bin_b2 = bin_b1(find(bin_b1 < max(T_mx)));

for i = 1:length(bin_a2)
    bin_e1{i,:} = TP_s(find(TP_s(:,1)>= bin_a2(i) & TP_s(:,1) <= bin_b2(i)),2);    % Bin of R_3     
end

Answer 1

[Steven Lord]

Rather than trying to implement the binning operation yourself, I recommend you call the histcounts function in a looping construct like while. Each call to histcounts in each loop iteration would use a different value for the BinWidth name-value argument until the results satisfy your requirements.

[Eli]

Hi Steven,

Could you provide an example code? As far as I am aware, we are unable to fix the bin width to 2 using histcounts. I wish to first fix the bin width as 2 and then only adjust if the bin does not contain 150 data points.

[Stephen23]

"As far as I am aware, we are unable to fix the bin width to 2 using histcounts. "

https://www.mathworks.com/help/matlab/ref/histcounts.html#buijikm-1-edges

[Steven Lord]

Make some sample data.

x = randn(1, 1e4);

Call the function.

[counts, edges] = histcounts(x, 'BinWidth', 2);

Check.

counts(1)

ans = 226

edges(1:2)

ans = 1×2

-4 -2

y = x(edges(1) <= x & x < edges(2))

y = 1×226

-2.0622 -2.9162 -2.1681 -2.3395 -2.3836 -2.3846 -2.5135 -2.2379 -2.2485 -2.5731 -2.6816 -2.1627 -2.8525 -2.1318 -2.1354 -2.0405 -2.3477 -2.2229 -2.0310 -2.3010 -2.5034 -2.5427 -2.4107 -2.4162 -2.0997 -2.5881 -2.1372 -2.1654 -2.7541 -2.6939

Note that y has counts(1) elements.