I have bi-weekly data for an event for which I am trying to build a forecasting model. When I plot the ACF and PACF, I get the following plots:
From what I understand, the plots show that the data are seasonal and seasonality has almost a fixed period of length 13 (as there are 13 bars in each block in the ACF plot). The data also seem to have a downward trend because of the auto correlation diminishes from left to right in the plot. My questions are:
- Am I interpreting the plots correctly?
- What types of models should I try with such data?
I have already tried
HoltWinters() from the
forecast package without much success. Any guidance is appreciated! Thanks!