How to use PDF to differentiate between signals?

Hi, everone I am trying to differentiate between three types of signals. And I want to use Probability Density Function (PDF) to do that. My signals are a non-stationary time serious type and they are a quite large (around 6 Millions data). I am still ABC in matlab. Therefore, Any help will be highly appreciated

[Muammer]

Could you please write The PDF code that I can use to differentiate between these signals

[Shashank Prasanna]

I can't write the code, since I am not really sure if what I suggested below is what you want. I urge you to read the documentation on the tests below and decide if that's what you want. If so, there are plenty of examples in the documentation page on how to use them. I am updating my post with the documentation links.

Answer 1

[Shashank Prasanna]

This is a generic question. So I would start with first visualizing the data by plotting its histogram using the HIST function.

From here you can either fit a specific distribution to your data, if you have reason to believe that it comes from a certain PDF, or you can fit it with a non-parametric PDF such as KSDENSITY function.

The final step is for you to compare distributions. This can be done in a number of ways. A popular method is the Kolmogorov-Smirnov test which can be called using kstest. You can directly provide the data to KSTEST function without fitting.

[Shashank Prasanna]

To compare densities: http://www.mathworks.com/help/stats/kstest2.html

To fit densities when distribution is unknown: http://www.mathworks.com/help/stats/ksdensity.html

[Muammer]

First of all, Many thanks for your help. Secondly, the following links are M.files for the signals I am talking about. I want to find a way of how to differentiate between them.It's not necessarily to stick with the PDF function, but any function that can do such. Your help will be highly appreciated. https://www.dropbox.com/s/5zwl2pi53coc9b2/Bu.mat https://www.dropbox.com/s/3xk21hjxudbmrfj/SttoSL.mat https://www.dropbox.com/s/dj7nzs6s5z7nklz/St.mat https://www.dropbox.com/s/s8fxcj8dcfoefej/SltoBu.mat https://www.dropbox.com/s/tzpteulmf8h2lgx/Sl.mat

[Shashank Prasanna]

Then your best bet is to start with KSTEST2 in the first link i gave above. That said there are too many ways to compare data or signals.

[Muammer]

I couldn't find a suitable way from those you mentioned. This is because I am still ABC in matlab. So,could you please give me a hand and guide me to the function that differentiates between those signals.

[Shashank Prasanna]

Ok, so first navigate to this link: http://www.mathworks.com/help/stats/kstest2.html Now scroll down to the example section. The first and only example on that page creates two vectors x and y and compares them. The null hypothesis is that they two vectors come from the same PDF If the output h=1 you are atleast 95% sure that they are different. Smaller the p value the more extreme the case and the more confident you are that they are different. (this is statistics not matlab). Read everything in the whole link. Copy past the code in MATLAB and run it with your data. If you still have questions please be very specific with your questions.