A system and method for optimising fertilisation

Abstract

There is disclosed a method for an agricultural fertilisation scheme and system for implementing the method. The method comprises collecting agricultural multidimensional data on farms, processing the collected data by dividing it into a training set and a test set, using logistic regression to develop a classification module for assessing the necessity of fertilisation in a pre-determined period of time, using an elastic net regression to construct a regression module that predicts the required fertilisation amount for the current pre-determined period of time, sequentially linking multiple classification and regression modules each tailored for predicting the application quantities of a fertiliser, and outputting the fertilisation amounts for multiple fertilisers. The predicted fertilisation amounts are validated using the test set. The validated fertilisation amounts are then implemented. Fig. 5