基于HP滤波分解的ARMA+BPNN的人民币汇率短期预测
|
摘要
汇率是影响国际经济的一个重要变量,对汇率进行准确的预测能有效地指导经济活动。运用时间序列相关理论,以美元兑人民币的月度数据为样本,运用HP滤波分解分析,将原始数据分解成平稳的趋势序列和方差时变的波动序列;对趋势序列建立ARMA模型并进行预测,对非线性的波动序列建立神经网络模型进行预测,然后两项预测值整合得到样本序列的预测值,并与单一的模型进行对比研究。结果表明,HP滤波分解法考虑了时间序列的波动性,在此分解基础上进行建模,而模型的短期预测效果更好。
Exchange rate is an important variable that affects the international economy.So accurate forecasting of exchange rates can effectively guide economic activity.In this paper,we use the theory of time series correlation,using the monthly data of USD/RMB as a sample.In the paper,we begin with HP filter to decompose exchange rate data series into the trend and fluctuation series,then establish ARMA model based on the trend series and BPNN model based on the fluctuation series to forecast,and integrate two predictors to obtain the prediction of the sample series,finally compare with a single model.The results show that the HP filter decomposition method which considers the volatility of the time series presents better forecasting effect.
Exchange rate is an important variable that affects the international economy.So accurate forecasting of exchange rates can effectively guide economic activity.In this paper,we use the theory of time series correlation,using the monthly data of USD/RMB as a sample.In the paper,we begin with HP filter to decompose exchange rate data series into the trend and fluctuation series,then establish ARMA model based on the trend series and BPNN model based on the fluctuation series to forecast,and integrate two predictors to obtain the prediction of the sample series,finally compare with a single model.The results show that the HP filter decomposition method which considers the volatility of the time series presents better forecasting effect.
引文
[1]谢赤.汇率预测与外汇干预研究[M].北京:科学出版社,2013.
[2]Jonathan D.,Cryer,Chan K.S.Time series analysis with applications in R[M].潘红宇,等,译.北京:机械工程出版社,2011.
[3]Maguire LP,Roche B,Mcginnity T M.Predicting a chaotic time series wing a fuzzy neural network[J].Information Sciences,1998,(112):125-136.
[4]Yu L.,Lai K.K,Wang S.Multistage RBF Neural Network Ensemble Learning for Exchange Rates Forecasting[J].Neuro computing,2008,(10).
[5]Thomas H.L.,Kyung D.N.,Robert E.M.Combining Foreign Exchange Rate Forecasts Using Neural Networks[J].Global Finance Journal,Spring-Summer,1998,(9):53-56.
[6]Hann T.H.,Steurer E.Much A do about Nothing Exchange Rates Forecasting:Neural Networks vs.Linear Models Using Monthly and Weekly Data[J].Neuro computing,1996,(10):23-27.
[7]Fang-Mei Tseng,Hsiao-Cheng Yub,Gwo-Hsiung Tzeng.Combining neural network mode with seasonal.time series ARIMA model[J].Technological Forecasting&Social Change,2002,(69):71-87.
[8]Lean Yu,Shouyang Wang,K.K.Lai.A novel nonlinear ensemble forecasting model incorporating GLAR and ANN for foreign exchange rates[J].Computers&Operations Research,2005,(32):2523-2541.
[9]熊志斌.ARIMA融合神经网络的人民币汇率预测模型研究[J].数量经济技术经济研究,2011,(6):64-76.
[10]彭兆祺,孙超.基于HP滤波分析方法的我国经济增长研究[J].山西财经大学学报,2011,(1):56-61.
[11]汤铎铎.三种频率选择滤波及其在中国的应用[J].数量经济技术经济研究,2007,(9):144-156.
[12]Weijun Xu,Ren Gu,Youzhu Liu,Yongwu Dai.Forecasting energy consumption using a new GM-ARMA model based on HP filter:The case of Guangdong Province of China[J].Economic Modelling,2015,(45):127-135.
[13]马慧慧,等.Eviews统计分析与应用:第3版[M].北京:电子工业出版社,2016.
[14]吴玉霞,温欣.基于ARIMA模型的短期股票价格预测[J].方法与应用,2016,(23).446.
[15]王钰,等.基于改进BP神经网络的预测模型及其应用[J].计算机测量与工程,2015,(1):39-42.
[16]许少华,宋美玲,许辰,等.一种基于混合误差梯度下降算法的过程神经网络训练[J].东北石油大学学报,2014,(4):92-96.
[17]曹广喜,曹杰,徐龙炳.双长记忆GARCH族模型的预测能力比较研究——基于沪深股市数据的实证分析[J].中国管理科学,2012,(20):41-49.
[18]惠晓峰,柳鸿生,胡伟,何丹青.基于时间序列GARCH模型的人民币汇率预测[J].金融研究,2003,(5):56.
[19]宋博,陈万义.基于HP滤波和ARMA-GARCH模型的人民币汇率趋势预测[J].数学的实践与认识,2017,(1):70-80.
[2]Jonathan D.,Cryer,Chan K.S.Time series analysis with applications in R[M].潘红宇,等,译.北京:机械工程出版社,2011.
[3]Maguire LP,Roche B,Mcginnity T M.Predicting a chaotic time series wing a fuzzy neural network[J].Information Sciences,1998,(112):125-136.
[4]Yu L.,Lai K.K,Wang S.Multistage RBF Neural Network Ensemble Learning for Exchange Rates Forecasting[J].Neuro computing,2008,(10).
[5]Thomas H.L.,Kyung D.N.,Robert E.M.Combining Foreign Exchange Rate Forecasts Using Neural Networks[J].Global Finance Journal,Spring-Summer,1998,(9):53-56.
[6]Hann T.H.,Steurer E.Much A do about Nothing Exchange Rates Forecasting:Neural Networks vs.Linear Models Using Monthly and Weekly Data[J].Neuro computing,1996,(10):23-27.
[7]Fang-Mei Tseng,Hsiao-Cheng Yub,Gwo-Hsiung Tzeng.Combining neural network mode with seasonal.time series ARIMA model[J].Technological Forecasting&Social Change,2002,(69):71-87.
[8]Lean Yu,Shouyang Wang,K.K.Lai.A novel nonlinear ensemble forecasting model incorporating GLAR and ANN for foreign exchange rates[J].Computers&Operations Research,2005,(32):2523-2541.
[9]熊志斌.ARIMA融合神经网络的人民币汇率预测模型研究[J].数量经济技术经济研究,2011,(6):64-76.
[10]彭兆祺,孙超.基于HP滤波分析方法的我国经济增长研究[J].山西财经大学学报,2011,(1):56-61.
[11]汤铎铎.三种频率选择滤波及其在中国的应用[J].数量经济技术经济研究,2007,(9):144-156.
[12]Weijun Xu,Ren Gu,Youzhu Liu,Yongwu Dai.Forecasting energy consumption using a new GM-ARMA model based on HP filter:The case of Guangdong Province of China[J].Economic Modelling,2015,(45):127-135.
[13]马慧慧,等.Eviews统计分析与应用:第3版[M].北京:电子工业出版社,2016.
[14]吴玉霞,温欣.基于ARIMA模型的短期股票价格预测[J].方法与应用,2016,(23).446.
[15]王钰,等.基于改进BP神经网络的预测模型及其应用[J].计算机测量与工程,2015,(1):39-42.
[16]许少华,宋美玲,许辰,等.一种基于混合误差梯度下降算法的过程神经网络训练[J].东北石油大学学报,2014,(4):92-96.
[17]曹广喜,曹杰,徐龙炳.双长记忆GARCH族模型的预测能力比较研究——基于沪深股市数据的实证分析[J].中国管理科学,2012,(20):41-49.
[18]惠晓峰,柳鸿生,胡伟,何丹青.基于时间序列GARCH模型的人民币汇率预测[J].金融研究,2003,(5):56.
[19]宋博,陈万义.基于HP滤波和ARMA-GARCH模型的人民币汇率趋势预测[J].数学的实践与认识,2017,(1):70-80.