Analysis of physicochemical and phytonutrients properties of bastard oleaster fruits and its mass prediction using artificial neural network model
Raju Sasikumar, Kambhampati Vivek, Govindasamy Kadirvel, Amit K. Jaiswal
Abstract
Bastard oleaster (Elaeagnus latifolia) is highly nutritious, rich in various vitamins (A, C, E) and minerals (P, K, Ca, Mg, Na), yet remains underutilized despite its immense potential. This study focused on the physicochemical and thermal properties of the fruit, and also aimed at accurately predicting its mass. A four-layered back-propagation artificial neural network (ANN) algorithm demonstrated excellent predictive capabilities, achieving the highest R2 values in training (0.9789) and testing (0.9901), with minimal error using the Log/Tan/Tan activation function trained with Trainlm. Additionally, the research assessed antinutritional properties and employed UPLC to identify and quantify phenolic compounds in E. latifolia. Rutin (11.02 mg/100 g) and epigallocatechin (7.56 mg/100 g) were identified as the most prominent compounds. Further HR-LC/MS analysis revealed specific functional compounds such as rhoifolin, which has antioxidant properties, and dihydrodeoxy streptomycin with antibiotic potential. The bioactive compounds like scopolamine, meperidine, and dextroamphetamine were also identified.