摘要
由于月降水量时间序列含有大量噪声,并表现出明显的混沌特性,现有预测模型均存在一定程度的不足.基于混沌理论的小波分析-VOLTERRA级数自适应(WA-VOLTERRA)耦合预测模型,在对月降水量时间序列进行混沌特性识别的基础上,先用小波分析对月降水序列进行时频分解,再分别对各频率分量进行相空间重构并用3阶VOLTERRA级数自适应模型建模预测,最后综合得到原始序列的预测值.以相近区域杭州市和南通市的月降水序列为例,并通过与小波分析-支持向量机(WA-SVM)模型进行比较,发现该模型具有较强的适用性和更高的预测精度.
To address the inefficiency of exsiting prediction models of monthly precipitation time series due to large amount of noises and obvious characteristics of chaos, a coupling model is proposed in this study, which takes full advantages of wavelet analysis and VOLTERRA adaptive model. The monthly precipitation time series is firstly mapped into several time-frequency domains, and then a third-order VOLTERRA adaptive model is established for each domain based on the phase-space recon- struction. The final forecasting results are the algebraic sums of all the forecasted components obtained by respective VOLT- ERRA adaptive model corresponding to different time-frequency domains. An experiment has been conducted by applying dif- ferent models to estimate the monthly precipitation time series in Hangzhou and Nantong, and the comparison of the data ob- tained by the conventional model with the results obtained using wavelet analysis and support vector machine （WA-SVM） cou- pling prediction model confirms that this new WA-VOLTERRA coupling method can achieve higher accuracy. The new model offers a new approach for monthly precipitation forecasting.
出处
《地球科学（中国地质大学学报）》
EI
CAS
CSCD
北大核心
2014年第3期368-374,共7页
Earth Science-Journal of China University of Geosciences
基金
国家自然科学基金项目(No.40801213)
