I am going to generate an ARIMA(0, 1, 1)
series with 60 observations. I want them to be strictly positive for all moving average parameters (ma
) while the series still follows an ARIMA(0, 1, 1)
process. Can I do that using arima.sim
?
arima.sim(n=100, model=list(ma=-0.9, order=c(0, 1, 1)))
This command give me both positive and negative values.
The sample space of an ARIMA process is the whole real line, so it is impossible to guarantee that the simulated values will be positive.
You could just add a constant to all values to make them positive. Or could you take the exponential of the simulated values.