Application of ANN for assessment of protein content, fat content and rancidity indicators as affected by the storage conditions and Rhizopertha dominica infestation in Pearl millet
摘要
Insect infestation contributes to substantial qualitative and quantitative deterioration in pearl millet grains in various storage regimes. This research examined the effect of different stages of Rhizopertha dominica infestation (1st day of infestation, egg, larvae, pupa and adult stages) at different storage temperature (15, 25 and 35 °C) on the quality deterioration of pearl millet grains. The impact of temperature was significant (p ≤ 0.05) on the crude protein and fat content. Pearl millet grains’ free fatty acid value (FFA), peroxide value (PV), and acid value (AV) were utilised as quality degradation indicators. Increase in storage temperature has seen an increase in the crude protein content, FFA, AV, PV and decreased crude fat content. The deterioration was severe at the adult stage of infestation, where protein content, FFA value, AV, and PV increased by 1.2 times, 2.5 times, 2 times, and 3–5 times, respectively, over the range of temperature under this investigation. On the other hand, fat content decreased by 1 factor for the same condition. The values of FFA, AV, PV, crude fat and protein content of the infested pearl millet grains were predicted using the artificial neural network (ANN) modelling based on the backpropagation method of Levenberg-Marquardt, with high R2 (0.983), and lower root mean square error (0.0053), which indicated the best fit of the developed model.