Divide et impera: How disentangling common and distinctive variability in multiset data analysis can aid industrial process troubleshooting and understanding

Abstract

[EN] The possibility of addressing the problem of process troubleshooting and understanding by modelling common and distinctive sources of variation (factorsorcomponents) underlying two sets of measurements was explored in a real-world industrial case study. The used strategy includes a novel approach to systematically detect the number of common and distinctive components. An extension of this strategy for the analysis of a larger number of data blocks, which allows the comparison of data from multiple processing units, is also discussed.