**AAS 205th Meeting, 9-13 January 2005**

*Session 153 Computation, Data Handling, Image Analysis*

Poster, Thursday, January 13, 2005, 9:20am-4:00pm, Exhibit Hall
Previous |
Session 153 |
Next

## [153.01] Bayesian Modeling of Astronomical Objects: Application to Planetary Nebulae

*K. A. Huyser, K. H. Knuth (NASA Ames), A. R. Hajian (USNO)*

In the course of understanding data acquired from
astronomical observations, astronomers often construct a
model of the object based on both the known physics and
acquired data. The construction of a model can become quite
complicated in terms of both the description of the object
and the amount of data for which the model must account. The
latter is becoming a more pressing issue as astronomers are
faced with increasingly large datasets.

The technique of Bayesian modeling proceeds as follows.
Given a parametric model of the object, one generates a
synthetic dataset from hypothesized values of the object's
model parameters. The synthetic data are compared to the
original data using the methods of probability theory. The
comparison results in the probability that these particular
parameter values could have given rise to the observed data.
The probabilities are then used to update the model
parameters. By repeating this sequence and iterating, one
may automate the process of estimating model parameters from
data.

The Bayesian methodology enables automation of parameter
estimation for large datasets once synthetic data generation
is accomplished. We describe the important concepts behind
Bayesian modeling and illustrate the procedures with an
application to Hubble Space Telescope images of planetary
nebulae.

Previous |
Session 153 |
Next

Bulletin of the American Astronomical Society, **36** 5

© 2004. The American Astronomical Society.