The paper presents an approach to time-series forecasting, which the reviewers thought is interesting and sound. Some of the concerns on the experimental side raised by the reviewers were successfully addressed during the rebuttal.