Probability Density Function (PDF) Estimator (V2.3)

An estimate of probability density function of the given random data with bounded support.

Description

This function provides an estimate of probability density function for a given random data (an 1-d vector). The estimation can be done with a specified number of intervals, and bandwidth. Without any output, the function will plot the probability density function. A few examples are included to show how to use the function and its output.

It also includes bounded support. For some problems, bounded support is important. For example, many physical problems require positive variables. This function is able to estimate PDF for such problems as well if the lower and upper bounds are specified.

The new version includes cdf and inverse cdf estimation.

Acknowledgements

The author wishes to acknowledge the following in the creation of this submission:
PDFPLOT
This submission has inspired the following:
Bivariant Gaussian Kernel Density Estimation, Update PDF Estimation

• MATLAB release: MATLAB 7.9 (2009b)
• HTML Files: Gaussian Kernel Density Estimation Demonstration
• Other Files: gkdeb.m, ; gkdetest.m, ; html/gkdetest_01.png, ; html/gkdetest_02.png, ; html/gkdetest_03.png, ; html/gkdetest_04.png, ; html/gkdetest_05.png, ; html/gkdetest_06.png, ; license.txt

[Jesper Henriksen]

I havent tried this code, but it sounds very similar to matlabs ksdensity

[Dimitri Shvorob]

It's an ad hoc smoother:
p.f(2:end-1)=p.f(2:end-1)*0.5+(p.f(1:end-2)+p.f(3:end))*0.25;
One can search FEX for 'kernel density estimator'.

[Yi Cao]

The file has been upgraded to use Gaussian kernel smoothing estimation algorithm.

[ricardo lusternik]

 

[V. Poor]