摘要
根据外汇汇率数据的特点,尤其是数据之间的依赖关系这一重要特性,并结合汇率预测要求快且准的原则,以数理统计学为基础,采用时间序列分析中的ARMA方法。通过数据的平稳性分析、模型的识别、参数估计、模型诊断以及最后预测精度的分析,建立了英镑兑美元汇率的ARMA预测模型并验证了其有效性。
According to the characteristic of the foreign exchange rate data, especially the reliance relation between the data, combining the exchange rate prediction requirement and based on the mathematical statistics the ARMA method in the time series analysis is adopted. Through analysis of the steadiness of the data, the discernment of the model, the parameter estimate; the model diagnosis and the analysis of the prediction precision the validity of the ARMA modeling method is verified.
出处
《成都信息工程学院学报》
2006年第6期914-919,共6页
Journal of Chengdu University of Information Technology
