I have a time series graph that is segmented into a few parts based on the maintenance day. You can think of it like vertical lines appearing out of the x axis which symbolize maintenance at the date. I would like to find out the relationship between the segmentation to predict what would happen to the signal after a maintenance day. Is there any way to tackle that problem? How should i solve it?
Considering that you are trying to predict values in the future according to a set of edges on you graph, you can do this in several ways. Humans predict some events of the future looking at the past, and you can model this with traditional time series forecasting methods, such as ARMA. You can build time series aggregating values for each segmentation and apply your forecasting.