Hi I have a couple of questions regarding an ARIMA(1,1,1) model
1. We have non-stationary data and want to estimate an ARIMA(1,1,1). Do we need to difference the series before running the model, or is this what happens when we run the model? Does D1. mean that the variable is differenced once?
2. When we run an ARMA(1,1) with differenced data and an ARIMA(1,1,1) with non-differenced data, on the same time series, the ARIMA uses much less observations. Why is this?
The picture above is an ARMA(1,1) with differenced data where there are 7054 observations, while the one below is an ARIMA(1,1,1) without differenced data and there are only 5560 observations.
1. We have non-stationary data and want to estimate an ARIMA(1,1,1). Do we need to difference the series before running the model, or is this what happens when we run the model? Does D1. mean that the variable is differenced once?
2. When we run an ARMA(1,1) with differenced data and an ARIMA(1,1,1) with non-differenced data, on the same time series, the ARIMA uses much less observations. Why is this?
The picture above is an ARMA(1,1) with differenced data where there are 7054 observations, while the one below is an ARIMA(1,1,1) without differenced data and there are only 5560 observations.
Attached Files
