What equivalent MATLAB commands plots PDF of data like 'hist' command plots histogram of data.

What command does the ploting of the PDF of a data in similar way as 'hist' command plots the histogram? I have a set of data from 1 to 500. My data ranges between -2 and 2. I want to plot the pdf such that the x-axis contains my data (which is -2 to 2 and not 1 to 500 ) while the y-axis is the PDF. When done, this should look similiar to the the result obtained in the 'hist' command.

NOTE: I do not want the result of the 'hist' command and the pdf on the same graph.

Answer 1

[Star Strider]

See if the histfit function does what you want.

.

[ojonugwa adukwu]

Thanks. It nearly does but two things: (1) I do not want the histogram in the plot; just the plot of the pdf only. (2) I will like the plot to touch the x-axis and not have an 'offset'.

[Star Strider]

That was not part of the original Question.

Without knowing more about the data, I can only suggest instead to use fitdist and normpdf.

data = 0.5*randn(50,1);
stats = fitdist(data, 'Normal')

stats = 
  NormalDistribution

  Normal distribution
       mu = 0.0155155   [-0.130784, 0.161815]
    sigma =  0.514782   [0.430015, 0.641487]

x = linspace(-4*stats.sigma+stats.mu, 4*stats.sigma+stats.mu, 150);
pdfv = normpdf(x, stats.mu, stats.sigma);

figure
plot(x, pdfv)
grid

Experiment to get results for the appropriate data.

.

[ojonugwa adukwu]

It does it. Thanks so much.

Meaning there is no single command like hist that plots the PDF of the data directly then.

[Star Strider]

As always, my pleasure!

Correct. There is no one command that will estimate the parameters and plot the distribution, at least that does it in the way you want. The handle of histfit has two components, the first being the handle to the bar object and the second being the handle to the line object, so in that sense, it is close. However it will not plot it exactly as you want, with the limits you require, so to get those requires the approach I use in my previous Comment.

.

Answer 2

[Ive J]

Something like this should work:

data = normrnd(10,1,100,1);
pd = fitdist(data, 'Normal');
x_values = 5:0.1:15;
y_values = pdf(pd, x_values);
plot(x_values, y_values, 'LineWidth', 2)

[ojonugwa adukwu]

Ok. I saw that in the documentation for histfit. Thanks