An Upgraded Interpolator of the Radial Basis Function Network for Spectral Calculation Based on Empirical Stellar Spectral Library

Stellar population synthesis is an important method in galaxy and star-cluster studies. In stellar population synthesis models, a stellar spectral library is necessary for the integrated spectra of the stellar population. Usually, the stellar spectral library is utilized for the transformation between the stellar atmospheric parameters and stellar spectrum. The empirical stellar spectral library has distinct advantages over the theoretical library. However, for the empirical spectral library, the distribution of stars is irregular in the stellar atmospheric parameter space, which makes the traditional interpolator difficult to get accurate results. In this work, we will provide an improved radial basis function interpolator which is implemented to obtain the interpolated stellar spectra based on the empirical stellar spectral library. For this interpolator, we use the relation between the standard variance sigma in the Gaussian radial basis function and the density distribution of stars in the stellar atmospheric parameter space to give the prior constraint on this sigma. Moreover, we also consider the anisotropic radial basis function by the advantage of the local dispersion of stars in the stellar atmospheric parameter space. Furthermore, we use the empirical stellar spectral library MILES to test this interpolator. On the whole, the interpolator has a good performance except for the edge of the low-temperature region. At last, we compare this interpolator with our work in 2018, and the interpolation result shows an obvious improvement. Users can apply this interpolator to retrieve the interpolated spectra based on the stellar spectral library quickly and easily.