The Application of Chemometric Methods in the Production of Enzymes Through Solid State Fermentation Uses the Artificial Neural Network—a Review

In the last decade, different multivariate statistical techniques have been applied to assist enzymatic production by microorganisms through solid state fermentation (SSF). The optimization of fermentative parameters such as temperature, time, pH, unit, aeration, spore concentration, and microbial strain significantly interfere in the process of enzymatic secretion by microorganisms. The advantage in using these statistical models is the reduction in the number of experiments, which provides savings in operational terms, in addition to the possibility of investigating the possible synergistic interactions between the fermentative parameters defined in the process. Statistical techniques such as central compound, Box-Behnken, Doehlert, and mix planning are limited to the experimental domain defined by the researcher, while the use of artificial neural networks (ANN), a tool based on artificial intelligence, eliminates this limitation and provides a mathematical model of the experiment. This review demonstrates the application of ANN for modeling experiments in SSF and its versatility to hybridize to different experiment optimization techniques. Thus, it is noticeable that the artificial neural network is a computational tool with the potential for replacing conventional statistical techniques, in addition to overcoming the limitations of these techniques, since ANN has the ability to extrapolate the experimental domain.