jovanajovicbonnet
commented
6 years ago Hi, I think you don't need to decompose a time series into its component if you don't want to. As far as I understood, decomposing will allow you to see all components of a signal separately and to visually check how your trend, seasonality (if you have it) and your noise look like
It's not very clear to me why we need to "decompose" the time series. I understand that it can show a plot with 4 rows but I'm not sure what I'm supposed to be mindful of when looking at the decomposition plots. Can anyone share the value of this? Thank you!