Development of Prediction Model for Chemicals in Fresh Fruits Using Artificial Neural Network

The objective of this investigation was to construct an artificial neural network (ANN) prediction model for chemicals like anthocyanin, titratable acidity, total solids soluble (TSS),vitamin C, titratable/TSS, and overall carotenoids levels of peach fruit employing surface color quantities, single fruit mass, liquid quantity, and sphericity percentage. In the initial hidden layer, an ANN framework with 6 inputs and fifteen neurons was built to predict 6 chemical compositional variables. Sensitivity testing found that liquid quantity was the most essential factor for determining titratable acidity,vitamin C, and titratable/TSS acidity. Furthermore, sphericity contributes 23.7% to anthocyanin and 24.0% to the overall carotenoids. Also, the color on TSS prediction had the largest contributing proportion of 20.8% when contrasted to the other characteristics. Chroma accounted for all parameters at varying levels ranging from 5.2 to 19.3%. Also, fruit mass attributed to every parameter at varying rates ranging from 16.6 to 23.4%. The ANN prediction approach represents a viable instrument for predicting the chemical compositional values of peach fruits at certain intended limits.