一种基于季节指数的灰色马尔科夫气温预测模型  

A Temperature Forecasting Model of Grey-Markov Based on Seasonal Index

作　　者：周华任[1] 李浩然[2] 孙学金[2] 颜万祥 

机构地区：[1]解放军理工大学理学院 [2]解放军理工大学气象海洋学院 [3]91079部队

出　　处：《数学的实践与认识》2016年第4期167-173,共7页Mathematics in Practice and Theory

摘　　要：提出一种根据气温历史数据的年际周期性和季节性变化规律建立的基于季节指数的灰色-马尔科夫气温预测模型.模型将纵向与横向分析相结合方法运用到气温预报之中,通过季节指数修正气温的横向季节性变化,再用灰色模型进行预测,最后通过马尔科夫进行误差修正.实例运用中,对广州市的2000年月平均气温进行预测,在与历史数据的对比中表明,模型预测结果较为准确,可靠性较好.并讨论说明该模型也可推广到其他具有周期特征的非平稳时间序列的预测中,并大大提高预测精度.According to the analysis of the inter-cyclical and seasonal variation of historical temperature data,a grey-Markovian model based on seasonal index for temperature forecasting is put forward.Attention was concentrated on the combination of horizontal and vertical analysis.Firstly the horizontal seasonal variation was adjusted by the seasonal index.Then the preliminary forecasting da-to was obtained by the grey system.Finally,the Markov model was introduced to correct the error of the preliminary data.The forecasts of monthly average temperatures Guangzhou in 2000 show that this model is more accurate and reliable.It shows that this model can be generalized to improve the forecasting precision of other non-stationary time series.

关 键 词：气温预测 季节指数 灰色理论 马尔科夫模型 

分 类 号：P457.3[天文地球—大气科学及气象学]