A Bayesian inference based two-stage support vector regression framework for soft sensor development in batch bioprocesses

• Published in: Computers & chemical engineering Vol. 41; pp. 134 - 144
• Main Author: Yu, Jie
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 11.06.2012; Elsevier
• Subjects: Applied sciences; Batch process; Bayesian analysis; Bayesian inference; Biological and medical sciences; Bioreactors; Biotechnology; Calibration; Chemical engineering; Computer science; control theory; systems; Control theory. Systems; Data processing. List processing. Character string processing; Exact sciences and technology; Fed-batch penicillin cultivation process; Fundamental and applied biological sciences. Psychology; Inference; Mathematical analysis; Mathematical models; Memory organisation. Data processing; Methods. Procedures. Technologies; Modelling and identification; Process and measurement uncertainty; Reactors; Regression; Sensors; Soft sensor; Software; Support vector regression; Various methods and equipments; Vectors (mathematics); Support vector regression; Fed-batch penicillin cultivation process; Soft sensor; Bayesian inference; Process and measurement uncertainty; Batch process; Input output; Level measurement; Modeling; Chemical reactor; Uncertain system; Penicillin; Vector support machine; System identification; Posterior distribution; Bayes estimation; Measurement sensor; Regression analysis; Experimental study; Bioreactor; Posterior probability; Data fusion; Alignment defect; Semicontinuous; Biased estimation

► A new Bayesian inference based two-stage support vector regression approach. ► Strong ability to identify and reconcile biased measurements under uncertainty. ► High accuracy and robustness for soft sensor predictions in batch bioprocess. ► Improved performance than SVR method in handling process...