基于加权的滑动平均—马尔科夫预测模型及其应用

降水量预测是雨水潜力化计算的首要环节.但由于降水过程存在高度的不确定性和随机性,很难用物理成因方法来确定某一时段降水量的准确值.因此可采用基于概率论和随机过程理论的马尔可夫模型进行预测[13],采用加权的方法,并通过对降水序列的滑动平均处理,降低序列的随机性,提高预测的准确性.以简阳1953-2004年年降水资料为分析对象进行实例分析,其中应用2006-2009年降水资料作模型检验,并对2010年-2005年年降水量作模型预测.结果表明:应用加权的滑动平均马尔科夫预测模型进行降水量预测是可行的,而且意义明确,计算简便,预测精度较高.

Markov prediction model and its application basedon the weighted sliding average

The precipitation forecast is the primary link in rainwater potential calculation. However, there is a high degree of uncertainty and randomness in precipitation process, it is difficult to determine the exact value of the precipitation in certain period of time with physical methods.Therefore Markov model can be used to predict based on probability theory and random process theory through using the method of weighted and sliding average treatment of precipitation sequence so as to reduce the randomness of the sequence and improve the prediction accuracy.Case study made the analysis of Jianyang precipitation data from 1953to 2004in which application of precipatition data from 2006to 2009was used to to do model checking. Therefore, model predicts applying this model of annual precipitation from 2010to 2015was done. The results shows that precipitation forecast is feasible applicating weighted moving average Markov prediction model. And the method is meaningful with simple calculation and high prediction accuracy.