Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 0711.4465

A method for exploiting domain information in astrophysical parameter estimation

C. A. L. Bailer-Jones
View original source ↗Read source document ↗Export metadata ↓

Abstract

I outline a method for estimating astrophysical parameters (APs) from multidimensional data. It is a supervised method based on matching observed data (e.g. a spectrum) to a grid of pre-labelled templates. However, unlike standard machine learning methods such as ANNs, SVMs or k-nn, this algorithm explicitly uses domain information to better weight each data dimension in the estimation. Specifically, it uses the sensitivity of each measured variable to each AP to perform a local, iterative interpolation of the grid. It avoids both the non-uniqueness problem of global regression as well as the grid resolution limitation of nearest neighbours.

Explore related subjects

astro-ph

Keep this discovery

BibTeXRIS

C. A. L. Bailer-Jones. 2007-11-28. A method for exploiting domain information in astrophysical parameter estimation. https://arxiv.org/abs/0711.4465

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2010

Circumstellar water vapour in M-type AGB stars: Radiative transfer models, abundances and predictions for HIFI

Aims: By performing a detailed radiative transfer analysis, we determine fractional abundances of circumstellar H2O in the envelopes around six M-type asymptotic giant branch stars. The models are also used to predict H2O spectral line emission for the upcoming Herschel/HIFI mission. Methods: We use Infrared space observatory long wavelength spectrometer spectra to constrain the circumstellar fractional abundance distribution of ortho-H2O, using a non-local thermal equilibrium, and non-local, radiative transfer code based on the accelerated lambda iteration formalism. The mass-loss rates and kinetic temperature structures for the sample stars are determined through radiative transfer modelling of CO line emission based on the Monte-Carlo method. The density and temperature profiles of the circumstellar dust grains are determined through spectral energy distribution modelling using the publicly available code Dusty. Results: The determined ortho-H2O abundances lie between 1e-4 and 1.5e-3 relative to H2, with the exception of WX Psc, which has a much lower estimated ortho-H2O abundance of only 2e-6, possibly indicating H_2O adsorption onto dust grains or recent mass-loss-rate modulations. The estimated abundances are uncertain by, at best, a factor of a few. Conclusions: The high water abundance found for the majority of the sources suggests that either the `normal' chemical processes are very effective in producing H2O, or else non-local thermal equilibrium atmospheric chemistry, grain surface reactions, or a release of H_2O (e.g. from icy bodies like Kuiper belt objects) play a role. We provide predictions for ortho-H2O lines in the spectral window of Herschel/HIFI.

astro-ph↗
arXiv2010

CMB Anisotropies and Inflation from Non-Standard Spinors

The apparent alignment of the cosmic microwave background multipoles on large scales challenges the standard cosmological model. Scalar field inflation is isotropic and cannot account for the observed alignment. We explore the imprints, a non-standard spinor driven inflation would leave on the cosmic microwave background anisotropies. We show it is natural to expect an anisotropic inflationary expansion of the Universe which has the effect of suppressing the low multipole amplitude of the primordial power spectrum, while at the same time to provide the usual inflationary features.

astro-ph↗