加权马尔可夫链预测模型的优化方法及应用

Research on an Improved Method of Weighted Markov Forecasting Model and Its Application in Predicting the Precipitation State

借鉴自回归模型,采取添加随机项的方法对加权马尔可夫链预测模型进行优化,并把优化后的模型应用于郑州市年降雨量的预测,通过与原模型和自回归模型预测结果的对比分析发现,优化后的模型弥补了原模型预测结果缺乏波动性的不足,同时使预测结果不仅在精度上有一定的提高,而且在趋势上更接近于实测值.

Through drawing lessons from AR model, the weighted method of Markov model by adding stochastic disturbance term on it is optimized, which is applied to a real instance. And the result tells that the improved Markov model overcomes the shortcoming of original Markov preferably, and the new model not only improves the forecasting veracity but also tend to become measured value, which provides an effective thinking and method to make weighted Markov model applied to the prediction of future precipitation.