How to display probabilities calculated for each data point

Dear Support,

I have a data sample x=(x1,....,xn)

suppose I want to fit negative binomial distribution to the data.

ie. X =NB(r,p)

However, for each data point in the sample, how do I display the probabilities calculated at each data point using Matlab.

Thanks in advance.

Silas

[Image Analyst]

I don't even know what that means. Your data is either true or false (1 or 0). You can have a PDF for the whole group of data that gives the likelihood that a particular element will be true or false, but I don't know what it means when you want to click on an element and then get the probability for just that element, as if it's different than for any other element or the set of data as a whole. Please explain by you think that each element has a probability function rather than being just a data sample.

[Silas Adiko]

Hi Shashank,

May be I was not explaining myself properly so let me put it another way:

Supposing I have a raw sample data and trying to fit a distribution. After estimating the parameters for different likely probability distributions, I wanted to calculate the Expected as against the Observed from data. This can be done recursively though But I think it is the probabilities needed at each data point that enables the Expected values to be calculated. So how do I code and display this e.g. for Negative Binomial or Poisson distributions.

[Shashank Prasanna]

This is still not clear, the expected value of NB is pr/(1-p) is this what you want? What do you plan to do recursively? Probability for each data is obtained by using the CDF as I mentioned. Please clarify your required approach and what how you want to do this.

Please use the comment option under the answer to reply to my answer instead of the question here.

Answer 1

[Shashank Prasanna]

Looks like you are interested in the negative binomial CDF. For your data X and for a given r and p maybe estimated from nbinfit you can use the CDF as described in the doc:

http://www.mathworks.com/help/stats/nbincdf.html