Data-Driven Prediction of Nitrogen Stress and Crop Yield in a Maize -Wheat Cropping System

Abstract

This study uses predictive analytics to deal with nitrogen stress and get the best crop returns in a system that grows both wheat and corn. We measured the amount of nitrogen (N) in the leaves along with a number of important farming factors using modern plant monitors like the GreenSeeker and SPAD meter. Different field studies gave us a range of nitrogen treatment rates for wheat (0 to 240 kg N/ha) and maize (40 to 300 kg N/ha). Machine learning methods, like Random Forest and Support Vector Machines, were used to create models that can accurately predict crop yields based on nitrogen doses, Normalised Difference Vegetation Index (NDVI) values, SPAD values, and direct measurements of leaf-N content. The results from pair plot studies showed strong positive links between leaf-N content, NDVI, SPAD values, and the rates of N application. These connections support the accuracy of the prediction models. Notably, the model that was tuned for wheat showed that it could be used with other crops by correctly predicting corn yields. It was amazing how well the prediction models worked. Random Forest and Support Vector Machines were able to get 85.71% and 100% accuracy on the original dataset, respectively.