Predictive Modeling of Agricultural Water Footprints in Iraq Using Machine Learning

Abstract

The research used machine learning and data mining to improve the use of water in agriculture in Iraq, assessing the water footprints of crops. Datasets from 4TU.ResearchData and Google Earth Engine were merged. The datasets were integrated into classification or regression algorithms to predict either crops locally planted or imported. The algorithms that used SVM and Random Forest achieved high accuracy successfully, which can help with the management of sustainable water and the selection of crops in regions with limited water resources.