I was wondering if anyone could help me to carry out prediction using a dynlm model. I seem to have trouble as there are two lagged components within the model. The model I've fitted is:
library(dynlm)
mod <- dynlm(x ~ 0 + trend + seasonal1 + seasonal2 + lag(x,-1) + lag(x,-2))
where x is a known time series of 100 values and seasonal1/seasonal2 are monthly/weekly seasonal components respectively. Trend is a linear trend component here given simply by:
trend <- 1:100
