Github user mengxr commented on the pull request: https://github.com/apache/spark/pull/3519#issuecomment-71615301 @zapletal-martin Thanks for the update! I made another pass. Besides inline comments, I have some high-level comments: 1. Should `PAV` take `isotonic` as a parameter? The implementation could be simpler if it only handles the isotonic case. Then the antitonic case could be done by flipping the sign of `x` first and then running isotonic regression. 2. I feel that we could compress the data. For each segment, we don't need to remember all points, but just the predictions on its end points. The solution will be a piecewise linear function.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org