Finding the probability distribution of a dataset

[Sujeeth Kumaravel]

Hello,
I have a dataset of 100 numbers. The numbers seem to be random. I wanna find the probability distribution of the dataset. How can I do it mathematically? I would really appreciate a reply.
Thank you

Sujeeth Kumaravel

 

[ledzep]

What kind of numbers do you have? Continuous, counts, binary?
You can look at the histogram and see if it normal or skewed...

 

[Sujeeth Kumaravel]

I have decimal numbers as data. Is there any mathematical method other than plotting histograms.
Thank you.

 

[Dason]

What you're asking to do is quite broad. Is it possible to choose an exact probability distribution? No - that wouldn't be possible. If you have a list of possible forms of probability distributions then you can choose between them. You can create a density estimate for the probability distribution which might in turn help you decide if you want to try to model it using a parametric form or you could just keep the density estimate as your estimate of the probability distribution directly.

But there are literally an uncountable number of possible probability distributions.

 

[Sujeeth Kumaravel]

Thank you. Now that you have suggested me to find the density estimate for the data, how do I find one.

 

[Dason]

Well a histogram is a very rough density estimate. A nicer method is kernel density estimation. Mosts stats packages implement these easily.

 

[catbelize]

If you have access to minitab, there is an option under 'stats' to use their 'Individual Distribution Identification' to test how different distributions fit your current data

 

[Sujeeth Kumaravel]

Thank you.