Fused stage-wise lasso – A waveband selection algorithm for spectroscopic data
|B. Zauner, T. Natschläger. Fused stage-wise lasso – A waveband selection algorithm for spectroscopic data. 11, 2013.|
A critical step in the analysis of spectroscopic data is the choice of an appropriate variable / wavenumber selection used for chemometric modeling. Regression methods such as PLS or PCR may theoretically be able to deal with a very large number of – possibly correlated – variables. In practice, however, it is widely known that variable selection can improve prediction performance, yield simpler models and simplify model interpretation.