Design Space Definition Using a Variational Bayes' Approximation

  • Type:
    Conference Presentation
  • Conference Type:
    AIChE Annual Meeting
  • Presentation Date:
    October 19, 2011
  • Skill Level:
  • PDHs:

Share This Post:

You will be able to download and print a certificate for PDH credits once the content has been viewed. If you have already viewed this content, please click here to login.

The design space may be interpreted as the constrained region of the key manufacturing parameters space which provides assurance of quality of a drug given that manufacturing occurs in this region. The metric used to represent this assurance should not be deterministic, rather it should be stated in probabilistic terms. In this work, we report on the use of Bayesian methods to develop a suitable risk metric based on statistical models of the manufacturing processes and product properties. The Bayesian inference is carried out to determine the posterior distribution of the probability of the product meeting quality specifications. Here, we propose an alternative optimization based procedure, the variational Bayes’ approximation, to obtain the posterior distribution. A sequential methodology is proposed to deal with non-linear models and the consideration of covariance. The results are compared to the widely used but computationally intensive Markov Chain Monte Carlo method. The proposed approached is illustrated with information drawn from a QbD study on Gabapentin.



Do you already own this?

Log In for instructions on accessing this content.


AIChE Member Credits 0.5
AIChE Members $15.00
AIChE Undergraduate Student Members Free
AIChE Graduate Student Members Free
Non-Members $25.00