Effects of application of biostimulants and growth regulators on root traits and analysis of root architecture with machine learning in Lilium

Apart from the roots, the bulb is the most important organ for plant development of the lily plant. In this experiment, the effects of mycorrhizal, vermicompost and promalin applications on the root architecture of Lilium oriental hybrid ‘Adelante’, a bulbous plant, were to be investigated. It was found that the effect of the treatments on root length (128.6 cm), root surface area (8 cm3), number of tips (111.5), number of forks (354.4) and number of crossings (86.2) was lower than that of the control. In terms of root volume, the applications of vermicompost (3 cm2) and promalin (3 cm3) were the most effective. The most effective application on root diameter (3.5 mm) was promalin. In conclusion, the effect on mycorrhizal root development was lower than the control but higher than other applications. In addition, machine learning (ML) algorithms, including linear regression (LR), sequential minimal optimisation for regression (SMOreg), Gaussian process (GP) and artificial neural network-based multilayer perceptron (ANN-based MLP), were used in the study. The input variables were evaluated for modelling and predicting root traits. The performance values of the ML algorithms were noted in the following order: LR > SMOreg > GP > MLP. These results have important implications for the prediction of root growth in lily crops.