Currently, when we traverse a diagonal, it may be best to restart the distance (or covariance) calculation after we pass a region in the time series where the time series signal dropped. I forget how this is currently being handled so we should investigate whether this is a problem.
Currently, when we traverse a diagonal, it may be best to restart the distance (or covariance) calculation after we pass a region in the time series where the time series signal dropped. I forget how this is currently being handled so we should investigate whether this is a problem.