Elastic net with Monte Carlo sampling for data-based modeling in biopharmaceutical manufacturing facilities. (2nd September 2015)

• Record Type: Journal Article
• Authors: Severson, Kristen; VanAntwerp, Jeremy G.; Natarajan, Venkatesh; Antoniou, Chris; Thömmes, Jörg; Braatz, Richard D.
• Is Part Of: Computers & chemical engineering. Volume 80(2015)
• Page Start: 30
• Page End: 36
• Publication Date: 2015-09-02
• Subjects: Data-based modeling -- Elastic net -- Lasso -- Monte Carlo methods -- Biopharmaceutical manufacturing -- Regularization methods; Chemical engineering -- Data processing -- Periodicals; 660.0285
• DOI: 10.1016/j.compchemeng.2015.05.006
• Languages: English
• ISSNs: 0098-1354

Abstract

Highlights: A statistical methodology is proposed for constructing process models. The methodology applies to heterogeneous data of limited availability. A regularization method, the elastic net, is combined with sampling techniques. The mathematical models have only a small number of input variables. This methodology is evaluated on an antibody manufacturing dataset. Abstract: Biopharmaceutical manufacturing involves multiple process steps that can be challenging to model. Oftentimes, operating conditions are studied in bench-scale experiments and then fixed to specific values during full-scale operations. This procedure limits the opportunity to tune process variables to correct for the effects of disturbances. Generating process models has the potential to increase the flexibility and controllability of the biomanufacturing processes. This article proposes a statistical modeling methodology to predict the outputs of biopharmaceutical operations. This methodology addresses two important challenging characteristics typical of data collected in the biopharmaceutical industry: limited data availability and data heterogeneity. Motivated by the final aim of control, regularization methods, specifically the elastic net, are combined with sampling techniques similar to the bootstrap to develop mathematical models that use only a small number of input variables. This methodology is evaluated on an antibody manufacturing dataset.