Grain Yield Prediction Based on the Metabolic Grey - Markov Integration Model

To improve the prediction accuracy of Chinese grain yield, a hybrid metabolic grey forecasting model based on the weighted Markov method is proposed. Since the grain yield is affected by many uncertain factors, the grain yield is predicted by the error amended metabolism grey model. Later, considering that the grain yield is also affected by the outputs over the years, the state transition probability matrixes are calculated, and the yield influence weights of the past years are decided. Lastly, combining the predicted yield and the influence weights, the final yield is corrected by recent years' yields, and the metabolic grey model is constructed. By using above procedures, the yields of 2016 to 2020 are predicted based on the data of 2005-2015, the results show that the forecasting error is less than 2% for all predicted years, and the mean error for 5 years achieves to 1.04%, which can be used to predict the grain yield accurately in the medium and short term.