Hi all,
I am modelling a time series with missing data.
*Q1)* However, I am not sure if I should use the next *graphics* to
understand my data:
*a)* ACF & PACF (original series)
*b)* ACF & PACF (residuals)
* *
*Q2)* I am using *tsdiag*, so I obtain a graphic with 3 plots: stand.
residuals vs time; acf for residuals; Ljung-Box for residuals (it is wrong
for residuals).
I know that using Box.test with type Ljung-Box, I can specify a correct df
to my estimated model (fitdf = p + q). So, I could do this test with
different lags, evaluate their significance, and then plot it. However, in
Box.test NA are not handled.
But, it is possible to do a Ljung-Box test with missing data [Stoffer &
Toloi, 1992. A note on the Ljung-Box-Pierce pormanteau statistic with
missing data].
*a)* Do you know any function to do a Ljung-Box test with NA?
*Q3) *In general, what (other?) tools do you recommend to use for time
series with missing data?
I had been using auto.arima and arima functions.
I don't want to do an interpolation.
Thanks in advance,
Cecilia
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
