基于不确定时间序列的外汇储量智能预测

Intelligent Prediction for Foreign Exchange Reserves Based on Uncertain Time Series

构建了一个关于外汇储量的不确定时间序列模型。通过最小二乘原理对不确定时间序列模型中的参数进行了估计,并对求得的不确定时间序列模型进行了残差分析和置信度预测。通过对不确定时间序列模型进行随机性测试证明了时间序列模型中的扰动变量是不稳定的,从而得出扰动变量实际上是不确定变量。此外,不确定假设检验的结果表明不确定时间序列模型可以很好地拟合外汇储量数据。

This article constructs an uncertain time series model about foreign exchange reserves.The parameters of the uncertain time series model are estimated by using the principle of least squares and residual analysis and confidence prediction of the obtained uncertain time series model are performed immediately.Finally,this article conducts a parameter estimation and a residual analysis of an uncertain time series model for actual data.Randomness tests are used to prove that the disturbance variables of the time series model are unstable.Thus the disturbance variables are actually uncertain variables.In addition,the result of the uncertain hypothesis test indicates that the uncertain time series model can well fit the foreign exchange reserve data.