Developing a Sustainable Machine Learning Model to Predict Crop Yield in the Gulf Countries

Crop yield prediction is one of the most challenging tasks in agriculture. It considered to play an important role and be essential step decision-making processes. The goal crop establish food availability for coming years, using different input variables associated with domain. This paper aims predict five Gulf countries’ crops: wheat, dates, watermelon, potatoes, maize (corn). Five independent were used develop a model, namely year, rainfall, pesticide, temperature changes, nitrogen (N) fertilizer; all these are calculated by year. Moreover, this research relied on widely machine learning models field prediction, which neural network model. model because it can complex relationships between dependent variables. To evaluate performance models, statistical evaluation metrics adopted, including mean square error (MSE), root-mean-square (RMSE), bias (MBE), Pearson’s correlation coefficient, determination coefficient. results showed that countries affected mainly four variables: pesticides, per average best RMSE R2 0.114 0.93, respectively. provides initial evidence regarding capability prediction.