Rstudio做时间序列ARIMA拟合时候发生错误

Karel.li

本人新手。。目前在看R IN ACTION
然后在 15.4 “ARIMA预测模型”一章中用forecas包+tseries包做Nile数据集的arima预测，出现以下错误：

> require(tseries)
> library(forecast)
> fit <- arima(Nile,order = c(3,0,3))
> fit

Call:
arima(x = Nile, order = c(3, 0, 3))

Coefficients:
          ar1     ar2     ar3     ma1      ma2      ma3  intercept
      -0.6320  0.5476  0.7582  1.0787  -0.0737  -0.6288   922.4686
s.e.   0.1287  0.1684  0.0902  0.1960   0.3377   0.1973    52.8730

sigma^2 estimated as 17934:  log likelihood = -633.65,  aic = 1281.31
> accuracy(fit)
              ME RMSE MAE MPE MAPE
Training set NaN  NaN NaN NaN  NaN
Warning message:
In trainingaccuracy(f, test, d, D) : test elements must be within sample
> forecast(fit,3)
Error in ts(x) : 'ts' object must have one or more observations



代码完全是按照书上来做的，可是出现这样子的结果，想问下各位是怎么回事？可以怎么解决？