具体问题如下：
predict(new2,h=1)
Error in predict.Arima(new2, h = 1) :
  'xreg' and 'newxreg' have different numbers of columns


其中new2是包含异常值一阶对数差分的自回归模型：
new2<-arima(log(xi),order=c(1,1,0),
xreg=data.frame(Y91=1*(seq(xi))==14,Y03=1*(seq(xi))==26)
)

其中xi是一组时间序列
xi<-c(112684.95,122633.57,140709.08,155620.56,187374.00,148485.00,161541.00
,151564.50,162629.60,181026.30,153102.30,138030.00,141614.40,87848.20
,101604.40,103972.00,100800.00,99087.12,101003.88,98084.00,74736.00,81718.00
,84543.20,84491.10,93931.20,73693.84,91502.51,96740.20,99582.20,136761.60
,139731.90,142476.00,142630.13,167341.94,187271.65,188851.64,209781.00)




其他人对此问题的看法
http://grokbase.com/t/r/r-help/133px65556/r-predict-arima-error-xreg-and-newxreg-have-different-numbers-of-columns
http://stackoverflow.com/questions/22208134/still-getting-error-after-set-newxreg-null-in-r
http://tolstoy.newcastle.edu.au/R/e10/help/10/05/3814.html