A COMPARISON OF SYNTHETIC AND OBSERVED SPECTRA FOR G-K DWARFS USING ARTIFICIAL NEURAL NETWORKS

A library of synthetic spectra, based on Kurucz model atmospheres, has been used to compare the spectroscopic observations in the wavelength range 4850-5384 Angstrom by using statistical and supervised artifici al neural network methods. The effective temperatures assigned by thes e methods for G-K dwarfs are compared with those given in Gray and Cor bally (1994) and found to be closely matching their calibration curve. This result provides a promising new technique for determination of f undamental stellar atmospheric parameters on the basis of comparison b etween the model generated synthetic spectra and observed stellar spec tra.