MATLAB Answers

How to plot a normal distribution graph to fit a bar graph?

184 views (last 30 days)
Rita
Rita on 24 Nov 2016
Commented: Star Strider on 24 Nov 2016
I have a bar graph which in the x-axis shows the edge centers and y-axis are N I would like to plot a normal distribution graph to fit the bar graph.Any suggestion?I know histfit but I have the N and edges only!Thanks

  2 Comments

Image Analyst
Image Analyst on 24 Nov 2016
So you want to fit the Normal distribution to the binned counts instead of the actual original data that you took the histogram of? Why? And why do you have only the counts and edges and not the original data? You had to have had the original data to get the counts, right?
Rita
Rita on 24 Nov 2016
Yes,I have calculated the percentage of the bins instead of counts for the yaxis.

Sign in to comment.

Accepted Answer

Star Strider
Star Strider on 24 Nov 2016
It’s easy enough to create your own version of histfit if you need to. See if this does what you want:
data = 0.5*randn(1,1000) + 2; % Create Data
mu = mean(data);
sd = std(data);
ndfcn = @(mu,sd,x) exp(-(x-mu).^2 ./ (2*sd^2)) /(sd*sqrt(2*pi)); % Standard Normal Distribution
[hc,edges] = histcounts(data, 25); % Histogram
ctrs = edges(1:length(edges)-1) + mean(diff(edges))/2; % Calculate Centres
ctrsx = linspace(min(ctrs), max(ctrs)); % High-Resolution Vector
sdnd = ndfcn(mu,sd,ctrsx); % Calculate Standard Normal Distribution
figure(1)
bar(ctrs, hc) % Plot Histogram
hold on
plot(ctrsx, sdnd*max(hc)/max(sdnd), '-r', 'LineWidth',2) % Plot Scaled Standard Normal Distribution
hold off
grid
Obviously, substitute your own data for my ‘data’ vector. I created mine to test my code.

More Answers (1)

Walter Roberson
Walter Roberson on 24 Nov 2016
I suggest reading the histfit() source and extracting the relevant portions of it -- using histfit() and figuring out a range with icdf and using pdf on the range

  0 Comments

Sign in to comment.

Tags