Predictive fertilization models for potato crops using machine learning techniques in Moroccan Gharb region

<span lang="EN-GB">Given the influence of several factors, including weather, soils, land management, genotypes, and severity pests diseases, prescribing adequate nutrient levels is difficult. A potato’s performance can be predicted using machine learning techniques in cases when there enough data. This study aimed to develop a highly precise model for determining optimal nitrogen, phosphorus, potassium required achieve both <br /> high-quality high-yield potato crops, taking into account impact various environmental factors such as soil type, management practices. We used 900 field experiments from Kaggle part data set. developed, evaluated, compared prediction models k-nearest neighbor (KNN), linear support vector (SVM), naive Bayes (NB) classifier, decision tree (DT) regressor, random forest (RF) eXtreme gradient boosting (XGBoost). measures mean average error (MAE), squared (MSE), R-Squared (RS), R<sup>2</sup>Root (RMSE) describe model’s mistakes capacity. It turned out that XGBoost has greatest R<sup>2</sup>, MSE MAE values. Overall, outperforms other models. In end, we suggested hardware implementation help farmers field.</span>