Abstract:
|
In this presentation, I will review some of contributions Cliff Spiegelman made to improving statistical analyses of chemical data for a variety of purposes. His ability to work with the chemists to understand the processes and the nature of the data made it interesting and fun to collaborate. A good example is receptor modeling in which we are looking at the mixture resolution problem, but without fixed chemical profiles and with much higher uncertainties in data compared to spectrochemical mixture resolutions where compounds have well defined and uniformly applicable spectral properties. Thus, interesting issues like uniqueness of solutions, constraints, numbers of resolvable components, and error estimation were the subject of deep and insightful conversations. His ability to bring rigorous statistics to the chemical field was greatly appreciated.
|