Mathematical Sciences: Multilinear Analysis of SpectroscopicData

这一统计学家和生物物理化学家之间的合作项目将改进和扩展光谱数据的分析。光谱学是对样品中颗粒的吸收或发射的测量。蛋白质的荧光光谱对许多不同的元素高度敏感，并且经常严重重叠。光谱数据的新特点是它的高信噪比。新的统计技术将利用数据的特殊特点，并针对感兴趣的具体问题进行分析。在标准因子分析中，假设响应变量的期望是未知参数的线性函数。光谱数据的性质使得有必要使用参数的非线性函数。

This collaborative project between a statistician and a biophysical chemist will refine and extend analyses of spectroscopic data. Spectroscopy is the measurement of the absorption or emission of particles from a specimen. The fluorescence spectra of proteins are highly sensitive to many different elements, and are frequently severely overlapping. The novel feature of the spectroscopic data is its high signal-to- noise ratio. The new statistical techniques will make use of the special features of the data and address the analysis to specific questions of interest. In standard factor analysis, the expectation of the response variable is assumed to be a linear function of unknown parameters. The nature of the spectroscopic data make it necessary to use non-linear functions of the parameters.