Interpreting autocorrelation plot from time series dataset

Question

I'm learning about time series forecasting. I have a dataset whose autocorrelation plot looks like this:

Here, the $X$ axis represents the lag in days, and the $Y$ axis represents the autocorrelation coefficient for each lag. I plotted the autocorrelation in hopes that it would show that my data set can be modeled according to either of the two models I've studied so far, which are the autoregressive model and the moving average model. That wasn't the case. However, looking at the plot, I see that:

The lags that are multiples of $7$ show a strong positive correlation.
The lags of the form $6 + 7k, \ 8 + 7k, $ with $ k \in \mathbb{N}$, show a weak positive correlation.
The lags of the form $2 + 7k,\ 3 + 7k, \ 4 + 7k, \ 5 + 7k$ with $ k \in \mathbb{N}$, show a medium negative correlation.

I was wondering if this pattern could be used in any way to forecast future data from this data set.

Please see our posts on seasonal time series. Because there are thousands of them, I focused the search and one of the top hits looks like it answers your question: https://stats.stackexchange.com/questions/144158. If not, please look through the remaining hits, including https://stats.stackexchange.com/questions/45191. — whuber, Mar 01 '23 at 23:35

Interpreting autocorrelation plot from time series dataset

0 Answers0