Stellar spectral classification is

Stellar spectral classification is one of the most fundamental tasks in survey astronomy. Many automatedclassification methods have been applied to spectral data. However, their main limitation is that themodel parameters must be tuned repeatedly to deal with different data sets. In this paper, we utilize theBayesian support vector machines (BSVM) to classify the spectral subclass data. Based on Gibbs sampling,BSVM can infer all model parameters adaptively according to different data sets, which allows us to cir-cumvent the time-consuming cross validation for penalty parameter. We explored different normalizationmethods for stellar spectral data, and the best one has been suggested in this study. Finally, experimentalresults on several stellar spectral subclass classification problems show that the BSVM model not onlypossesses good adaptability but also provides better prediction performance than traditional methods.