Prediction of the kind of sprouts of Cruciferae family based on artificial neural network analysis

The aim of this work was to show that artificial neural networks (ANNs) are a convenient tool for predicting the kind of sprouts originated from Cruciferae family. For this purpose, the known contents of bioactive compounds of small radish, radish, white mustard, and rapeseed seeds and sprouts were used for the prediction of the kind of a specific sprout. The input data reflected the contents of the following compounds in cruciferous seeds in the course of germination: soluble proteins (SP), ascorbic acid (AH2), total glucosinolates (GLS), reduced glutathione (GSH), and tocopherols (α-T, β-T, γ-T, δ-T), expressed in their biological activity. The ANN used was trained on the learning set. The ability of the utilised ANN to generalise the gained knowledge based on the learning set was verified by the validating and testing sets. The trained and validated ANN was able to classify, with complete accuracy, the kind of sprouts out of the four kinds used. It can be concluded that ANN can be used as a useful tool for determining the identity of cruciferous seeds and sprouts based on the determined levels of their bioactive components.