Introducing hidden Markov models to LAMOST automatic data processing

摘要

The LAMOST1,2 telescope is expected to have its first light in later of 2007. The 4-meter aperture and 4000-fiber feeding ablility will make it a powerful spectra sky survey instrument, as well a challenge to the mission of data processing and analysis. So far several statistical methods, mainly based on PCA, have been developed for spectra automatic classification and red shift measurement by a team of LAMOST3. Statistical methods of Hidden Markov Modelling have become popular in many area since 1990s, which are rich in mathematical structure and can form the theoretical basis for use in a wide range of applications, e.g. speech recognition and pattern recognition. No doubt they are prospective implements for automatic spectra processing and analysis. In this paper, I attempt to briefly introduce the theoretical aspects of this type of statistical modelling and show the possible applications in automatic spectra data processing and analysis.