Artificial neural networks (ANNs) were used in this study to assess the performance of Deep learning neural networks and Wide learning neural networks in predicting Chlorophytum borivilianum in vitro organogenesis. The effects of different compositions of macronutrients of MS media were analyzed for the micropropagation of Chlorophytum borivilianum using nodal explant to predict shoot organogenesis (outputs) by the application of different ANN architectures i.e., Wide neural network and Deep neural network. The number of hidden layers in a deep learning architecture is more than that of a WNN. The ANN architecture was trained using the Broyden-Fletcher-Goldfarb-Shanno (LBFGS) quasi-Newton algorithm. To determine which model was superior in predicting shoot organogenesis, we examined their respective accuracy percentages and R2 values (the coefficient of correlation). When comparing the Regression neural network with the Wide neural network model, the former achieved a greater accuracy percentage (99.99991022) when predicting the total number of shoots. The Shoot length was more accurately predicted (97.58222034) by the WNN model. There is a slight discrepancy between the accuracy percentage and the R2 value, but only at the decimal level. The results of nonparametric biological studies can be predicted with equal accuracy using either architecture.

Chlorophytum borivilianum, Deep learning, Wide neural network, Macronutrients, invitro Regeneration