Output details

Interpretation of non-linear empirical data-based process models using global sensitivity analysis

This article solves a pending-problem in statistical design of experiments and response surface methodology, where it is known that flexible non-linear models can help design better processes/products with fewer experiments. However, unlike linear regression, these complex non-linear models are difficult to interpret, in terms of how inputs affect outputs. Through the application of global sensitivity analysis, such interpretation becomes possible, and it will promote the uptake of the highly accurate non-linear models in improved process and product design.