Is there a way to generate poles of a system/ a transfer function using autoregressive model?

Hi,

I have some time series data from which I want to find the poles of the system. The transfer function is unknown.

I am using an autoregressive model. My aim is to find the poles of the system and determine which ones belong to the useful signal and which ones are noise components in order to get the required signal.

I have used the 'ar' command in matlab, also tried 'arburg' in matlab. I do get an answer, however, as far as I understand, I get the coefficients of the ar model from both. How/is there a command or a method to find the poles of the unknown system?

My code goes like this:

data = [72.7678 73.2033 72.8667 72.9444 73.2511 72.9144 73.1667 73.0656 72.8933 73.1922 73.1744]; 
% please note that there are about 400 elements in this array. I only copied some due to space requirements.
order = 9; 

m = ar(data,order);

or 

aa = arburg(data,order);

I get many values for m 1x0 idpoly. Here is a snapshot of the result.

Can someone help me if it is possible to find the poles and also how to?

Thanks.