摘要
为探求时序数据间的相关性与灰色模型的模拟预测效果之间的关系,提出了先对时序数据进行自相关分析,在判断自相关程度高低的基础上,建立等维灰数递补动态模拟预测模型.并分别对渭河流域的林家村站和华县站1983—2000年的年径流数据以及安徽省1989—2006年的工业和生活废水排放量数据进行了实例分析,发现自相关程度越高,精度越高.最后根据时序的自相关程度和精度检验的高低,给出了确定进行短期预测或者中短期预测的建议.
To relate time series autocorrelation and grey model predication effect, a scheme is put forward. Firstly, autocorrelation analysis is made for time series to obtain degree of autocorrelation. Then a dynamic forecasting model with reeursive compensation by grey numbers of identical dimensions is developed. The natural runoff for 1983 to 2000 from Linjiacun and Huaxian gaging station, and industrial and domestic wastewater discharge for 1989 to 2006 in Anhui province is used. It is found that higher the degree of autocorrelation, higher the precision. Based on autocorrelation analysis and precision inspection, suggestions are made in regards to choosing short-term predication or long-term predication.
出处
《北京师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第6期640-644,共5页
Journal of Beijing Normal University(Natural Science)
基金
国家科技支撑资助项目(2006BAD20B06)
关键词
自相关分析
等维灰数递补动态预测模型
精度检验
autocorrelation analysis
dynamic forecasting
recursive compensation
grey numbers
precision inspection
