The PLS method -- partial least squares projections to latent structures -- and its applications in industrial RDP (research, development, and production)

The chemometrics version of PLS was developed around 25 years ago to cope with and utilize the rapidly increasing volumes of data produced in chemical laboratories. Since then, the first simple two-block PLS has been extended to deal with non-linear relationships, drift in processes (adaptive PLS), dynamics, and with the situation with very many variables (hierarchical models). Starting from a few examples of some very complicated problems confronting us in RDP today, PLS and its extensions and generalizations will be discussed. This discussion includes the scalability of methods to increasing size of problems and data, the handling of noise and non-linearities, interpretability of results, and relative simplicity of use.