恒星大气物理参量的非参数估计方法

恒星大气物理参量（有效温度、表面重力、化学丰度）是导致恒星光谱差异的主要因素．恒星大气物理参量的自动测量是LAMOST等大规模巡天望远镜所产生的海量天体光谱数据自动处理中一个重要研究内容．针对测量大样本的恒星光谱数据估计每个恒星的大气物理参量，提出了一种基于变窗宽核函数的估计算法：变窗宽算法是对固定窗宽算法的改进，分为3个步骤：（1）将历史恒星光谱数据进行PCA处理，得到光谱的低维特征数据；（2）利用特征数据与其物理参数的对应关系，建立一种变窗宽的非参数估计模型；（3）利用该估计模型，直接计算待测恒星光谱的3个物理参量（有效温度、表面重力、金属丰度）．实验结果表明：该方法与固定窗宽估计模型以及在其他文献中报道的方法相比，具有较高的估计精度和鲁棒性．

Non-parameter Estimation Algorithm to Determine Stellar Fundamental Parameters

The three fundamental parameters of stellar atmosphere, which are the effective temperature, the surface gravity, and the metallic abundance, determine the continuum and spectral lines in the stellar spectrum. A new algorithm is proposed for estimating the three parameters in this paper. It is composed of three steps: (1) the sample data of stellar spectrum, whose three fundamental parameters are known, is reduced to a lower dimensional feature space using PCA (Pnincipal Component Analysis) technique; (2) a non-parameter estimator with variable window-width is set up from the correspondence between the feature data and their fundamental parameters data; (3) this estimator is used to compute the fundamental parameters of a stellar spectrum. Experiments indicate that the estimator is more efficient and robust than the fixed window-width estimator.