artificial neural networks (mlp and rbf) as tools for weight prediction of orchid synthetic seeds produced using an encapsulation set-up

The synthetic seed method refers to encapsulated plant parts and any meristematic tissue which can develop into plantlets under in-vitro or in-vivo conditions. various parameters and evaluating’ one-variable-at-a-time’ could be time-consuming, expensive, and inefficient. thus, the application of process modeling approaches including multi-layer perceptron (mlp) and the radial-basis function (rbf) can be required and beneficial for the prediction of synthetic seed weight. in the present study, two different types of artificial neural network (ann) algorithms, the mlp and rbf models, have been developed to predict the weight of phalanopsis orchid synthetic seed using an encapsulation set-up especially developed for this purpose. various topologies of ann were configured based on different concentrations of sodium alginate (3, 4, and 5 (w/v)), calcium chloride (100,125, and 150 (mm), and droplet falling height of sodium alginate (1, 1.5, and 2 cm) as input variables and the values of synthetic seed weights as output variable. results show that the rbf algorithm (r= 0.98 and sse= 0. 13× 10-3) outperformed the mlp algorithm (r = 0.91and sse= 0.14× 10-3) owing to its better ability for predicting capsule weight. the study has presented a machine learning-based approach for the classification of synthetic seeds. algorithms for extraction of capsule features have been developed, which are in turn used to train artificial neural network (ann) classifiers. the outputs of anns have been successfully applied to model the synthetic seeds production process indicating the appropriateness of the model equation in predicting orchid synthetic seed weight are mathematically combined.