A deep learning-based prediction model for soybean yield

The objective of the paper is to propose a state-of-the-art deep learning approach to crop yield prediction for soybean based specifically on artificial neural networks (ANNs) and convolutional neural networks (CNNs) using time-series meteorological data, soil data and the Enhanced Vegetation Index (EVI). The performance of the proposed model is to be compared to the random forest model, a standard model used for crop yield prediction. Furthermore, the influence of the input parameters on the yield are to be analyzed so as to determine their significance in yield prediction models.