purifier
Full Member level 4
I am a beginner in signal processing so I apologize in advance if I end up using the wrong terminology. My goal is to find out correlated events in a set of time series data (image attached below) that I have. I figured that using cross-correlation, I will be able to do this. In the first image (the time series plot), I can clearly see that the 3rd plot and the 10th plot obtained similar patterns around the same time. To understand this, what I did was to plot the cross-correlation function for the two time series (which is also attached next). The cross-correlation plot tells me that the two time series had a strong correlation with a lag of 5 days which is no doubt useful to me. However, I still don't quite understand how to mark the sections that had the strongest correlation i.e. in this example, I want to find out the period around the 600th mark that had the strongest correlation.
I am thinking that if I use a sliding window of say 20 days and compute cross-correlation for all the graphs, I will be able to achieve this but I was wondering if there are any standard techniques in signal processing to achieve this.
I am thinking that if I use a sliding window of say 20 days and compute cross-correlation for all the graphs, I will be able to achieve this but I was wondering if there are any standard techniques in signal processing to achieve this.