Finding distribution to which data belongs.

I think you can also look into Model Selection. Where you do a Bayesian analysis over different models to see which ones fits your data the best.

Andre Ng has a great couple of pages on Model Selection, and you can look for it on Bishop's book as well.

It's a bit more elegant solution that uses a Bayesian point of view.

A light introduction to this topic and considerations related to it is Huber's slides from a talk he gave on "Fitting Distributions to Data: Practical Issues in the Use of Probabilistic Risk Assessment" (PDF).

There is also a write up on some approaches on how to do this using R written by Ricci "Fitting Distributions with R" (PDF).

Bayesian approaches are very popular in machine learning. So you might consider a paper such as Gelman's "A Bayesian Formulation of Exploratory Data Analysis and Goodness-of-fit Testing" (PDF) a possible starting point as well.

There are numerous other exploratory techniques out there, but these should be suitable starting points that can lead to additional resources or questions. I hope this helps guide you a bit as to an approach. Otherwise feel free to ask a follow up and I (or I'm sure someone else around) will certainly be glad to do our best to help further. Good luck in your search!