how to get the arima model's variance of white noise's at each points in R?

Question

I have trained a ARIMA(1,0,2)(0,1,1)[7] just like that:(use R's function Arima{forecast} )

 >test.arima=Arima(x=tsx.rd_lm, order=c(1,0,2), seasonal=list(order=c(0,1,1),period=7),fixed=c(NA,0,NA,NA), lambda=0.9533674)

 >test.arima
Box Cox transformation: lambda= 0.9533674 

Coefficients:

     ar1  ma1     ma2     sma1
  0.6089    0  0.2314  -0.8650
s.e.  0.0426    0  0.0513   0.0383
sigma^2 estimated as 170690303:  log likelihood=-4908.65
AIC=9825.3   AICc=9825.43   BIC=9845.84

How ever, the sigma value {170690303} is terriable, so I hope get each points white noise (variance) to check. After all, maybe some xreg parameter could be find more effectivly. @ADD: a brief of the data:Named num [1:241] 732499 724785 717221 709805 702539 ... @ADD: how ever, test.arima doest't offer the names like a lm class in R. So, just like get the lm model's residuals at each point by residuals lm.my$residuals, How to get the arima class's white noise's at each points in R?

Thanks!

Answer 1

residuals(test.arima)

However, you should be aware that the size of $\sigma^2$ depends on the scale of the data. Divide your data by 1000, and your $\sigma^2$ value will be 170.690303.