Grain Yield Prediction Based on the Improved Unbiased Grey Markov Model

Grain yield is affected by a variety of complex factors, with large volatility and randomness. In order to improve the accuracy of grain yield prediction, this paper proposes a grain yield prediction method with improved unbiased grey Markov model. In the unbiased grey Markov model, after the average division of states is performed, the values within each state are nonlinear. This paper proposes to replace the method of taking the median of the first and last transitions in the unbiased Markov chain with the average for the calculation of state division and to correct the residual values of the prediction results by using the improved unbiased grey Markov model, in order to improve the accuracy of the predicted value of grain yield. Simulation experiments were conducted to compare the grey GM (1, 1) model, the unbiased grey GM (1, 1) model, the unbiased grey Markov model and the improved unbiased grey Markov model. The original grain output data for Chongqing from 2000 to 2022 were used for the validation analysis to compare the prediction accuracies of the four models. The results show that the prediction accuracy of the grey GM (1, 1) model and the unbiased grey GM (1, 1) model is basically the same, with an average error of 3.213%. The prediction accuracy of the unbiased grey Markov model is better, with an average error of 2.039%. The unbiased grey Markov model has the smallest prediction average error of 1.367%. Compared with the previous three models, the improved unbiased grey Markov model can further improve the prediction accuracy, which is suitable for medium- and long-term prediction and predicts the grain production data of Chongqing in the next eight years.