Model size determination using data analysis tools in the context of reaction network identification

Data dimensionality analysis or model size determination (identifying the true dimension of the underlying data variant space from experimental data) is a key step for the accurate identification of reaction networks and kinetic modeling of chemical systems. Here, several data analysis tools are compared regarding their relative performance in the identification of the model size from simulated noisy data concerning a challenging case study that presents low reaction extents. Depending on the characteristics of the data used, distinct choices of analytic methods can be suggested. The complementary application of different methods to reinforce the confidence of this analysis is also considered.