Predicting crop yields in Senegal using machine learning methods

Abstract Agriculture plays an important role in Senegalese economy and annual early warning predictions of crop yields are highly relevant the context climate change. In this study, we used three main machine learning methods (support vector machine, random forest, neural network) one multiple linear regression method, namely Least Absolute Shrinkage Selection Operator (LASSO), to predict food staple crops (peanut, maize, millet sorghum) 24 departments Senegal. Three combination predictors (climate data, vegetation data or a both) compare respective contribution statistical inputs predictive skill. Our results showed that with gives best performance. The prediction skill is obtained for peanut yield likely due its high sensitivity interannual variability. Although more research needed integrate study into operational framework, paper provides evidence promising performance methods. development operationalization such their integration systems could increase resilience Senegal change contribute security.