Soft Sensor Modeling Using Machine Learning for Fermentation Process

The aim of the present book has been to develop soft sensor solutions for upstream bioprocessing and demonstrate their usefulness in improving robustness and increasing the batch-to-batch reproducibility in bioprocesses. This book study encompasses the following objectives:- To propose and compare the performance of successive projection algorithm with grey relation analysis algorithm in terms of auxiliary variables selection; - To propose and compare the performance of SPA-GWO-SVR soft sensor model with SPA-SVR model in terms of accuracy, root mean square error, coefficient determination R2;- To propose exponential decreasing inertia weight strategy with PSO algorithm that exploits search space and thus by reducing large step lengths leads the PSO towards convergence to global optima; - To propose the fuzzy c-means clustering algorithm to cluster the sample data and compare the performances of the IPSO-LSSVM soft sensor model with standard PSO-LSSVM model on selected benchmarked regression datasets in terms of accuracy, mean square error, root mean square error, and mean absolute error.