Modeling of peach production energy using machine learning in Nazarabad township, Alborz province

Today, providing food security for the world's growing population by preserving the earth's resources & minimal environmental effects has become one of the basic & important challenges in sustainable agriculture, & the optimal use of resources is one of the main requirements of sustainable agriculture. In this study, the pattern of energy consumption during peach production, analysis & modeling of energy & performance of peach production in Nazarabad city was investigated. Data were collected through interviews & filling specialized questionnaires. The results showed that the total energy consumption & production were equal to 72716.83 & 5234.89 megajoules per hectare, respectively. Electricity was the most consumed input with a share of 59% of the total input energy. The indices of energy efficiency, energy efficiency, energy intensity and net energy were obtained as 0.07, (kg/MJ) 0.03, (MJ/Kg) 26.39 and (MJ) -67481 respectively. Modeling was done with three methods enhanced gradient regression, decision tree regression, and random forest regression, and RRMSE was -0.003, -0.0090, and -0.0091, and R2 was 0.98, 0.95, and 90, respectively was calculated. The results showed that the enhanced gradient method is able to accurately predict the values of the energy efficiency indices of peach production. The results showed that the energy efficiency and production can be predicted with high accuracy through the inputs of irrigation water, electricity, chemical and animal fertilizers, labor force, chemical poisons, diesel fuel and machines and machine learning method. Sensitivity analysis was performed with SHAP and the results showed that the most influential input in energy prediction was nitrogen fertilizer.