The Optimization of the Halophilic Cellulase Production: A 3-2-1 Multilayer Perceptron Artificial Neural Network Approach

Lignocellulose is one of the bio-resources available on the earth. It could be hydrolyzed into simple sugar. The previous study found that carboxymethylcellulose (CMC), FeSO4·7H2O, and NaCl are the significant mediums that influence the production of halophilic cellulase. Despite that, an appropriate method is deemed crucial from an industrial perspective to optimize halophilic cellulose production for cost-effectiveness. In this regard, the optimum halophilic cellulose production is determined from the best-so-far parameter of the three significant mediums. A data mining process using Multilayer Perceptron (MLP) based on the Artificial Neural Networks (ANN) method is developed to optimize the parameter from a set of experimental data. A 3-2-1 MLP network was constructed to learn the experimental data. As a result, the root squared error from the MLP is 0.0118 during the validation process. Subsequently, the MLP network was considered to determine the parameter of the significant medium and the production of halophilic cellulose. Consequently, this finding provides beneficial guidance for the manufacturer in the chemical industry to achieve efficient halophilic cellulose production.