Histogram with adjusted bins to gaussian
2 views (last 30 days)
Show older comments
I need to take the new histogram from this code and apply a gaussian but when I try to do so I get a gaussian over the unaltered histogram.
here is my code
[num,~,~]=xlsread('histogram2.xlsx');
d=num;
h = histogram(d,50)
%remove one count per bin for background estimate
h.BinCounts(h.BinCounts>0) = h.BinCounts(h.BinCounts>0)-1
h=histfit(d,50)
I dont know how to referance the histogram with the adjusted bins in the histfit function
I am on R2020a
0 Comments
Answers (1)
Tommy
on 16 May 2020
If you'd like to fit a histogram to a normal distribution but you don't know the underlying data (e.g. you've altered the histogram in some way), you could instead fit the bin centers and values to a Gaussian curve:
% create your histogram
[num,~,~]=xlsread('histogram2.xlsx');
d=num;
h = histogram(d,50)
h.BinCounts(h.BinCounts>0) = h.BinCounts(h.BinCounts>0)-1
% fit the Gaussian, based only on h
x = h.BinEdges(1:end-1)+h.BinWidth/2;
y = h.Values;
gaussian = @(mu, sig, scale, x) 1/(sig*sqrt(2*pi))*exp(-(((x-mu)/sig).^2)/2) * scale;
x0 = [mean(x), range(x), sum(h.Values*h.BinWidth)]; % guesses for [mu, sig, scale]
f = fit(x(:), y(:), gaussian, 'StartPoint', x0);
% plot the Gaussian on top of your histogram
hold on;
p = plot(f);
p.LineWidth = 2;
You can obtain the fit parameters with f.mu and f.sig. You may need to use a better starting point.
2 Comments
See Also
Categories
Find more on Histograms in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!