New Preprint about Unsupervised Machine Learning on Diverse Measurement Data

Already since a few years our research group is exploring approaches for the automated and – most importantly – unbiased analysis of experimental data. In particular Anton Vladyka and Maria El Abbassi have been driving forces in that respect.

My colleague Mickael Perrin has taken this further by developing a generalised approach for the analysis of various kinds of measurement data and performing an extensive benchmark of a large range of algorithms. Together, we have extended this to my particular field of interest, Raman spectroscopy.

Our discussion of the various methods for dimensionality reduction, clustering and cluster validation are available in the following preprint:
We are happy about any feedback!