Uploaded image for project: 'Public H2O 3'
  1. PUBDEV-3482

Supporting GLM binomial model to allow two arbitrary integer values

    Details

    • Type: New Feature
    • Status: Resolved
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 3.10.2.1
    • Component/s: None
    • Labels:
      None
    • CustomerVisible:
      No
    • Sprint:
      H2O Dev Sprint 34, H2O Dev Sprint 33, H2O Dev Sprint 34
    • AffectedContact:
      Steve Lavdas <steve.lavdas@nielsen.com>
    • AffectedCustomers:
    • Support Assessment:
      Data Science Issue
    • Customer Request Type:
      Support Incident

      Description

      • Added 8th DEC 2016 ----
        Requirement- Implement 3.2 section of attached PDF in GLM and GBM while GLM is first to deliver.

      Supporting “pseudo” logistic regression, by allowing GLM binomial model which currently supports 0/1 for the values of the target to allow two arbitrary integer values (for example -4,7).

      More Info: How to use GLM (or really classification models in general) without the restirction of the response being categorical (for example it can be -4 and 7). Sections 3.1 and 3.2 go into the detail. Please see the attached paper.
      TMLErare.pdf

        Attachments

          Activity

            People

            • Assignee:
              tomas Tomas Nykodym
              Reporter:
              avkash Avkash Chauhan
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: