Impact of measurement error in Bayesian design space determination for pharmaceutical processes

The aim of this thesis is to propose a systematic approach for the incorporation of measurement uncertainty in the Bayesian identification of the DS of a new pharmaceutical product. Specifically, the proposed approach extends a joint Bayesian/latent variable methodology for DS identification recently proposed by Bano et al. (2018). A step-by-step methodology is proposed to handle measurement uncertainty in the calibration dataset, and three case studies are used to test its performance.