We present a method for deriving stellar fundamental parameters. It is based on a regularized sliced inverse regression (RSIR). We first tested it on noisy synthetic spectra of A, F, G, and K-type stars, and inverted simultaneously their atmospheric fundamental parameters: Teff, log g, [M/H] and vsini. Different learning databases were calculated using a range of sampling in Teff, log g, vsini, and [M/H]. Combined with a principal component analysis (PCA) nearest neighbors (NN) search, the size of the learning database is reduced. A Tikhonov regularization is applied, given the ill-conditioning of SIR. For all spectral types, decreasing the size of the learning database allowed us to reach internal accuracies better than PCA-based NN-search using larger learning databases. For each analyzed parameter, we have reached internal errors that are smaller than the sampling step of the parameter. We have also applied the technique to a sample of observed FGK and A stars. For a selection of well studied stars, the inverted parameters are in agreement with the ones derived in previous studies. The RSIR inversion technique, complemented with PCA pre-processing proves to be efficient in estimating stellar parameters of A, F, G, and K stars.

We aim at finding the value of an explanatory variable, through its expression in a large data-vector, without knowing the link function between the explanatory variable and the data-space. Sliced Inverse Regression (SIR) method allows for the projection of a data-vector onto a subspace consistent with the explanatory variable variation. We suggest a method based on the SIR subspace, that gives the most efficient estimation of an unknown explanatory variable.

Our aim is to evaluate fundamental parameters from the analysis of the electromagnetic spectra of stars. We may use $10^3$-$10^5$ spectra; each spectrum being a vector with $10^2$-$10^4$ coordinates. We thus face the so-called "curse of dimensionality". We look for a method to reduce the size of this data-space, keeping only the most relevant information.As a reference method, we use principal component analysis (PCA) to reduce dimensionality. However, PCA is an unsupervised method, therefore its subspace was not consistent with the parameter. We thus tested a supervised method based on Sliced Inverse Regression (SIR), which provides a subspace consistent with the parameter. It also shares analogies with factorial discriminant analysis: the method slices the database along the parameter variation, and builds the subspace which maximizes the inter-slice variance, while standardizing the total projected variance of the data. Nevertheless the performances of SIR were not satisfying in standard usage, because of the non-monotonicity of the unknown function linking the data to the parameter and because of the noise propagation. We show that better performances can be achieved by selecting the most relevant directions for parameter inference. Preliminary tests are performed on synthetic pseudo-line profiles plus noise. Using one direction, we show that compared to PCA, the error associated with SIR is 50$\%$ smaller on a non-linear parameter, and 70$\%$ smaler on a linear parameter. Moreover, using a selected direction, the error is 80$\%$ smaller for a non-linear parameter, and 95$\%$ smaller for a linear parameter.

We present an automated procedure that derives simultaneously the effective temperature $T_{eff}$, the surface gravity logg, the metallicity [Fe/H], and the equatorial projected rotational velocity vsini for "normal" A and Am stars. The procedure is based on the principal component analysis inversion method of Paletou et al. (2015a). A sample of 322 high resolution spectra of F0-B9 stars, retrieved from the Polarbase, SOPHIE, and ELODIE databases, were used to test this technique with real data. We have selected the spectral region from 4400-5000\AA\ as it contains many metallic lines and the Balmer H$\beta$ line. Using 3 datasets at resolving powers of R=42000, 65000 and 76000, about 6.6x$10^6$ synthetic spectra were calculated to build a large learning database. The Online Power Iteration algorithm was applied to these learning datasets to estimate the principal components (PC). The projection of spectra onto the few PCs offered an efficient comparison metric in a low dimensional space. The spectra of the well known A0- and A1-type stars, Vega and Sirius A, were used as control spectra in the three databases. Spectra of other well known A-type stars were also employed in order to characterize the accuracy of the inversion technique. All observational spectra were inverted and atmospheric parameters derived. After removal of a few outliers, the PCA-inversion method appears to be very efficient in determining $T_{eff}$, [Fe/H], and vsini for A/Am stars. The derived parameters agree very well with previous determinations. Using a statistical approach, deviations of around 150 K, 0.35 dex, 0.15 dex, and 2 km/s were found for $T_{eff}$, logg, [Fe/H], and vsini with respect to literature values for A-type stars. The PCA-inversion proves to be a very fast, practical, and reliable tool for estimating stellar parameters of FGK and A stars, and deriving effective temperatures of M stars.

We show how the range of application of the principal component analysis-based inversion method of Paletou et al. (2015) can be extended to late-type stars data. Besides being an extension of its original application domain, for FGK stars, we also used synthetic spectra for our learning database. We discuss our results on effective temperatures against previous evaluations made available from Vizier and Simbad services at CDS.

The general context of this study is the inversion of stellar fundamental parameters from high-resolution Echelle spectra. We aim indeed at developing a fast and reliable tool for the post-processing of spectra produced by Espadons and Narval spectropolarimeters. Our inversion tool relies on principal component analysis. It allows reduction of dimensionality and the definition of a specific metric for the search of nearest neighbours between an observed spectrum and a set of observed spectra taken from the Elodie stellar library. Effective temperature, surface gravity, total metallicity and projected rotational velocity are derived. Various tests presented in this study, and done from the sole information coming from a spectral band centered around the Mg I b-triplet and with spectra from FGK stars are very promising.