Plot multiple histogram within 1 function

Hi all! I made some histograms, but now I'm trying to make a function that plots multiple histograms at the same time. I want 1 plot with 4 overlapping histograms of different colors and a legend. Can someone please explain to me how to do this? I tried to make a loop (see bottom) but it doesn't seem to work.

histogram_color = [0 .7 .7];
histogram(error1,'normalization','pdf','edgecolor',histogram_color,'facecolor',histogram_color,'facealpha',.5)
x = -90:1:90;
f1 = fitdist(error1,'kernel');
pdf1 = pdf(f1,x);
l1 = line(x,pdf1,'linestyle','-','color',histogram_color,'linewidth',3);

histogram_color = [0 0 .7];
histogram(error2,'normalization','pdf','edgecolor',histogram_color,'facecolor',histogram_color,'facealpha',.5)
x = -90:1:90;
f2 = fitdist(error2,'kernel');
pdf2 = pdf(f2,x);
l2 = line(x,pdf2,'linestyle','-','color',histogram_color,'linewidth',3);

x = -90:1:90;
for i = 1:length(out)
histogram(out(i),'normalization','pdf','edgecolor',histogram_color,'facecolor',histogram_color,'facealpha',.5);
f(i) = fitdist(out(i),'kernel');
pdf(i) = pdf(f(i),x);
l(i) = line(x,pdf(i),'linestyle','-','color',histogram_color,'linewidth',3);
end

[Adam Danz]

You're missing "hold on"

Otherwise the 2nd histogram replaces the first.

Answer 1

[Sudhakar Shinde]

[~,ErrorLength]= size(out);
for i =1:ErrorLength
    hold on
    histogram(out(:,i),'normalization','pdf')
    hold off
end

Answer 2

[Steven Lord]

The "Plot Multiple Histograms" example on the documentation page shows how to superimpose two histogram plots using hold on.

Answer 3

[MadjeKoe]

S = load('/Users/mkoenraad/Desktop/Scriptie/exercise_2/simulated_serial_bias_dataset.mat');

res = S.res;

targets = res(:,7);

responses = res(:,8);

orx = [3,4,5,6];

tgt = [1,2,3,4];

num = numel(orx);

out = cell(1,num);

for k = 1:num

idx = targets==tgt(k);

resp = responses(idx);

ori = res(idx,orx(k));

err = resp - ori;

err(err<-90) = err(err<-90)+180;

err(err>90) = err(err>90)-180;

out{k} = err;

end

out = [out{:}];

ero1 = out(:,1);

ero2 = out(:,2);

ero3 = out(:,3);

ero4 = out(:,4);

[MadjeKoe]

Is this ok? I don't know how to only add my 'out' variable

[Sudhakar Shinde]

histogram_color = [0 .7 .7];
x = -90:1:90;
[~,ErrorLength]= size(out);
for i =1:ErrorLength
    figure(i)
    hold on
    histogram(out(:,i),'normalization','pdf','edgecolor',histogram_color,'facecolor',histogram_color,'facealpha',.5)'
    f = fitdist(out(:,i),'kernel');
    pdf = pdf(f,x);
    l = line(x,pdf,'linestyle','-','color',histogram_color,'linewidth',3);
    hold off
end

[Sudhakar Shinde]

yes. ok. Try this code at bottom. it works fine.

[MadjeKoe]

I keep on getting this message:

Unable to use a value of type prob.KernelDistribution as an index.

Error in forumans (line 35)

pdf = pdf(f,x);

I don't understand why because it worked before

[Sudhakar Shinde]

Try this code:

[~,ErrorLength]= size(out);
for i =1:ErrorLength
    hold on
    histogram(out(:,i),'normalization','pdf')
    hold off
end

This gives output on single graph with different histograms without using kernel. As i dont have toolbox i can not check for

f = fitdist(out(:,i),'kernel');

pdf = pdf(f,x);

[MadjeKoe]

Thank you so much!!!! This is exactly want I meant!

[Sudhakar Shinde]

Glad to help you.

[Adam Danz]

You only need to "hold on" once and that was mentioned in a comment under the question 13 hours prior to this sub-thread.

hold on %                   <---  Move here
[~,ErrorLength]= size(out);
for i =1:ErrorLength
    % hold on               <--- Remove 
    histogram(out(:,i),'normalization','pdf')
    % hold off              <--- remove or move out of the loop
end

[Adam Danz]

"I did not work."

I doubt it. It's the same solution as Sudhakar Shinde's and you indicated that it worked. There must have been some other problem (which is likely given the new propblem follwing "clear all").

" I tried 'clear all' & now my whole thing is not working anymore"

'clear all' is rarely necessary. You likely were using a global variable named x and now it's no longer defined. If that's the case, it s a good lesson to never use global variables. Now you need to figure out where x was defined.