Tobacco single leaf weight prediction method based on meteorological coupling soil fertility of big data

The invention discloses a big data-based meteorological coupling soil fertility tobacco single leaf weight prediction method. The method comprises the steps of obtaining meteorological data and soil fertility data in a prediction period; and inputting the meteorological data and the soil fertility data into a single-leaf prediction model, and calculating the single-leaf weight of the tobacco. According to the method, meteorological data and soil fertility data are combined, a Transform model effectively captures a long-distance dependency relationship between input features through a self-attention mechanism, the model can comprehensively understand and utilize interaction between meteorological data and soil data, compared with a traditional statistical prediction model, the method can more accurately predict the weight of a single tobacco leaf, and the method has the advantages that the method is simple and convenient to operate and high in practicability. Therefore, agricultural producers are helped to bet