Photosynthetic Rate Prediction Model of Golden Melon Plant (Cucumis melo L.) at Vegetative Phase in Greenhouse using Artificial Neural Networks

Abstract

The most critical parameter affecting plant growth is the photosynthetic rate. The parameter can be determined by measuring the rate of CO2 assimilation that occurs in plants. Developing a photosynthetic rate model can recommend proper cultivation maintenance in melon plants. Hence, the involvement of input parameters in the developed model affects the accuracy of the prediction. This study aims to develop an artificial neural networks (ANNs) prediction model of the photosynthetic rate of melon plants in the vegetative phase in the greenhouse based on seven environmental and growth parameters and find the best model structure. Model development uses artificial neural networks with several stages: data collection and pre-processing, model development with different input variations, model validation, and selection of the best scenario to predict photosynthetic rate. The results showed that five out of seven input parameters, i.e., air temperature, sunlight intensity, CO2 concentration, air humidity, and plant rows, in the model structure of five inputs, six hidden and one output were the best model scenarios with coefficient of determination (R2) and root mean square error (RMSE) of 0.986 and 0.420, respectively.