摘要
文章采用三种模型：自回归模型、ETS指数平滑状态空间模型和非参数方法中的局部多项式回归模型对人民币汇率进行点预测,在此基础上引入广义诱导有序加权调和平均算子（GIOWHA）的概念,以预测值和实际值的p次幂倒数误差的平方和最小为准则,构建基于GIOWHA算子的最优变权系数组合预测模型。结果表明GIOWHA组合预测模型具有更精准的预测能力。
This paper uses auto-regression model, ETS exponential smoothing model and local polynomial regression model to forecast the RMB exchange rate, on the basis of which generalized induction ordered weighted harmonic average operator（GIOWHA） is introduced to construct a GIOWHA-based optimal combination forecast model with variable weight coefficient, following the principle of minimizing the sum of squares of the P power reciprocal of error between the predicted value and the actual value.The study result shows that the GIOWHA combination forecast model has more accurate forecasting ability.
作者
丁小松
杨桂元
Ding Xiaosong;Yang Guiyuan(School of Statistics and Applied Mathematics, Anhui University of Finance ＆ Economics, Bengbu Anhui 233000, Chin)
出处
《统计与决策》
CSSCI
北大核心
2018年第12期9-13,共5页
Statistics & Decision
基金
国家社会科学基金资助项目(12BTJ008)
