Seminarium i matematisk statistik
fredagen den 11:e december 13.15 i  MH227

Anders Malmberg presenterar sitt examensarbete

Titel: Fluorescence rejection in Raman spectroscopy by 
       local polynomial kernel regression methods

Abstract:

The nonparametric kernel regression method has proven to
be a good choice when estimating regression functions and
their derivatives. Here, this method is used in order to
estimate Raman spectra. A local bandwidth selector is used
when estimating the measured summarised spectrum and when
estimating the fluorescence part of the spectrum the SiZer
method might be a useful tool. The localisation of the peaks
is decided studying pointwise confidence intervals for the
derivative.