摘要

Based on the usual chaotic-weighting first-rank local-region forecasting model, a modification is made for the predictor in matrix and vector simulation instead of single variant. Then the modified forecasting model is applied to predict silicon content in molten iron of the medium-sized blast furnace in China, and good results are obtained. Finally, we find that the value of Kolmogorov entropy has a great effect on the hitting accuracy of the prediction.