when I last looked into time series clustering, it seemed like other statistical techniques worked better for predicting failure modes. I still think there is m

when I last looked into time series clustering, it seemed like other statistical techniques worked better for predicting failure modes. I still think there is merit in motif recognition and rules based around that concept, but I haven't seen very much of interest outside of EKGs.