You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ignite.apache.org by "Aleksey Zinoviev (JIRA)" <ji...@apache.org> on 2018/09/13 15:45:00 UTC

[jira] [Created] (IGNITE-9587) [ML] Umbrella ticket: Handle different labels in training data and handle unknown labels in test or updated training data correctly

Aleksey Zinoviev created IGNITE-9587:
----------------------------------------

             Summary: [ML] Umbrella ticket: Handle different labels in training data and handle unknown labels in test or updated training data correctly
                 Key: IGNITE-9587
                 URL: https://issues.apache.org/jira/browse/IGNITE-9587
             Project: Ignite
          Issue Type: New Feature
            Reporter: Aleksey Zinoviev
            Assignee: Aleksey Zinoviev


The problem is that all algorithms of binary classification are ready to handle the datasets marked with 0/1 labels and predict 0/1 labels without especial mapping.

Also the algorithms don't handle situation with unknown labels during the updating and testing phases

Possible solution: it could be stored in context of ML training



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)