Advanced Process Control for Production of Influenza Vaccine Shariah Compliance using Acutase in Batch Reactor

Abstract

Influenza viruses cause significant worldwide morbidity and mortality every year. Muslims all over the world are debating whether the vaccine qualifies as halal. When this topic is brought up again, some of them even took the chance of forgoing the vaccine. Muslim reluctance and rejection of vaccination is founded on presumptions and convictions that such a procedure and the substances employed are "doubtful (mashbooh)". Process control for the production of shariah compliance vaccine mainly focused on bioreactor to produce hemagglutinin (HA) using Acutase. The process control is carried out via artificial intelligence approach. Artificial neural network (ANN) was employed to create a shallow neural network to predict the HA yield using experimentation data sourced from a relatable journal, whereby the inputs were pH, temperature (^oC), pressure (bar), reaction time (day) and oxygen level. A performance comparison and evaluation was carried out to analyse and validate the Levenberg-Marquardt (LM) & Bayesian Regularisation (BR) based-ANN models by altering the number of neurons. Overall, LM generalised better than BR algorithm as all of LMANNs’ MSE values were kept low. It was determined that LMANN is suitable for the experimental database used to build the ANN models.