CIM Seminar with Ragnar Seton

Abstract: My research focuses on the properties and applications of stripline split-ring resonator micsoplasma sources. Using these lightweight, lower power sources for optical emission spectroscopy we have in previous studies demonstrated their usefulness in applications ranging from planetary exploration to blood gas measurements in neonatal care. In this talk I will give a short overview of the research topic and then focus on the data analysis involved in predicting individual species concentrations from the emission spectra of one of these microplasmas.
In particular, I will address the predictive performance of a bagging neural network and compare that to more traditionally used partial least squares methods using an experimentally obtained dataset of CO2-N2-Ar emission spectra.

Registration: please, click here to register