Model-driven and data-driven approaches in Plant yield Prediction
December 5th, 2016
In agriculture research, the Crops Yield Prediction (CYP) is always an important topic. Normally, the agricultural planners should estimate the yields for all crops before the planting season. However, it is always difficult to do such a job because the Crop Yield Prediction depends on a lot of interrelated factors, like their genotype or the environmental condition where they are planted. What’s more, the farmers’ decision has also great influence on the Crop Yield, such as land preparation, irrigation, sowing date or using fertilizers. Sometime, the market can also influence the farmer’s decision. Yield prediction traditionally has relied on farmers’ long-term experience for specific fields, crops and climate conditions, which can be inaccurate. Simple estimators, such as the average of several previous yields or the last obtained yield, are also used. Nevertheless, crop yield varies spatially and temporally with non-linearity, introducing large deviations from one year to another. Thus, more efficient methods have been developed, which can be classified as crop growth models and data-driven models. So in this presentation, we try to make a comparison of these two principal approaches for Crops Yield Prediction.