Machine Learning Models - Is There a Neighbor?

In one of the original research papers, there's a topic called time series chains. And really what this means is that if you analyze, you take your time serie an you compute the matrix profiles. The question that you can ask is, where is that nearest neighbor located? So if it happens to be located in the future, then for that subsequence, where’s that next nearest neighbors located? You can kind of have this chaining effect, right? Where you keep moving forward in time, across time. They demonstrated in once you compute these time series chains, their example that they used was looking at coals retail stores. It actually showed some not only seasonality,