Identifying most influencing input parameters for predicting Cereal production using an artificial neural network model

Abstract

Growth of populations and climate change are the most factors that affect agricultural production, particularly in developing countries. Several variables affect agriculture production, expressed by cereal production, prediction; therefore, identification of the most significant parameters for accurate prediction is an important research area. In the present study, artificial neural network (ANN) models with various combinations of input parameters were developed to determine the most significant parameters that influence the cereal production prediction in Lebanon- neighboring countries. To achieve this, urban population (UP), rural population (RP), urban population growth (UPG), rural population growth (RPG), population density (PD), agricultural land (AL), Agricultural nitrous oxide emissions (ANoE), Methane emissions (ME), Nitrous oxide emissions (NoE) and CO2 emissions (CO2E) were considered as input variables. The output variables were cereal production (CP) and land under cereal production (LUC). 35 ANNmodels were developed by varying the identified input parameters. Additionally, the coefficient of determination (R2) and root mean squared error (RMSE) were used to select the best predictive model. Out of the 35 ANNmodels, ANN-13, ANN-26, ANN-28, and ANN-33 have given the best prediction with the combinations of [PD and CO2E], [CO2E, ANoE, ME, NoE], [PD, CO2E, ANoE, ME, NoE] and [RPG, PD, CO2E, ANoE, ME, NoE], respectively. Consequently, the growth of populations and CO2 emissions are the most parameters that have influenced the prediction of CP and LUC.