T. F. Adepoju, O. Olawale & S. K. Layokun
An Artificial Neural Network (ANN) was engaged to optimize the effect of β-Cyclodextrin on the production of Phenyl methanol (PM) from biotransformation of benzaldehyde by free cells of yeast. In developing ANN model, performance of ANN is heavily influenced by its network structure, five-level-five-factors design was applied, which generate 50 experimental runs. The inputs for the ANNs are cell weight (wet. wt): X1, incubation time (min): X2, Acetaldehyde conc. (mg/100 ml): X3, benzaldehyde conc. (mg/100 ml): X4, and β-level (%): X5. The learning algorithms used was QP with MNFF, the transfer function was Tanh. Meanwhile, RMSE was determined to be 3.0739. The coefficient of determination R2 and the adj. R2 were found to be 0.99206 and 0.98419, respectively. It was observed that 900 (mg/100 ml) benzaldehyde with 1000 (µg/100 ml) acetaldehyde in the presence of 1.8% β-cyclodextrin gave the highest yield (351.5 mg/100 ml) of PM. Hence, it can be concluded that yeast (Saccharomyces cerevisae) can tolerate higher levels of acetaldehyde and benzaldehyde due to the effects of β-cyclodextrin.