Exploring the relationship between maize yield and climate big data on maize belt of northeast China

Data-driven agronomy is becoming popular by reducing the climate uncertainty in the era of agriculture big data. In this paper, many data mining approaches were used to extract embedded knowledge from climate variability and explore the relationship between NDVI data and maize yield. The results show that the Oct NDVI is important for Liaoning, while the Aug NDVI is important for Heilongjiang and Jilin. The approaches presented here were especially valuable when combined with key month forecasts. The Neural Networks, which could explain approximately a half of the yield variability, is suitable for Heilongjiang and Liaoning providence, while the bootstrap forest, which less than 50% explain ability, is suitable for Jilin.