Greenhouse Temperature and Humidity Estimator Design Based on Least Square SVM

Résumé

In this paper, a Least Squares Support Vector Machine (LS-SVM), as a modern approach to data analysis and modelling using the methods and tools of artificial intelligence, was employed to model and predict the temperature and relative humidity as the prime climatic parameters for the development and growth of the crop under a greenhouse. In the development of LS-SVM models, linear kernel, polynomial kernel and Gaussian radial basis function (RBF were applied to train the LS-SVM. For that, external climate variables and command inputs were considered as input variables to the model to get the estimated greenhouse temperature and relative humidity. The performances of the LS-SVM models, computed through the Root Mean Square Error (RMSE) and determination coefficient. The proper LS-SVM tuning parameters, the regularization parameter γ and the kernel function parameter σ were adjusted to get the best LSSVM outputs. The simulation results showed that RBF kernel boasted better precision and generalization than linear and polynomial kernels and thus able to provide precision for the relevant greenhouse climate parameters estimation.