You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@systemml.apache.org by "Deron Eriksson (JIRA)" <ji...@apache.org> on 2016/08/30 00:01:40 UTC

[jira] [Updated] (SYSTEMML-887) Automatic DataFrame type determination in MLContext

     [ https://issues.apache.org/jira/browse/SYSTEMML-887?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Deron Eriksson updated SYSTEMML-887:
------------------------------------
    Description: 
Currently the new MLContext API supports the following 4 DataFrame types:
1) ID column with doubles
2) No ID column with doubles
3) ID column with vector
4) No ID column with vector

When feeding a DataFrame input into the new MLContext API, the DataFrame type can be specified as a MatrixFormat.

For usability, the DataFrame type should also be dynamically determined from the data structure and data itself.

  was:
Currently the new MLContext API supports the following 4 DataFrame types:
1) ID column with doubles
2) No ID column with doubles
3) ID column with vector
4) No ID column with vector

When feeding a DataFrame input into the new MLContext API, the DataFrame type can be specified as a MatrixFormat.

For usability, the DataFrame type can also be dynamically determined from the data structure and data itself.


> Automatic DataFrame type determination in MLContext
> ---------------------------------------------------
>
>                 Key: SYSTEMML-887
>                 URL: https://issues.apache.org/jira/browse/SYSTEMML-887
>             Project: SystemML
>          Issue Type: Task
>          Components: APIs
>            Reporter: Deron Eriksson
>            Assignee: Deron Eriksson
>
> Currently the new MLContext API supports the following 4 DataFrame types:
> 1) ID column with doubles
> 2) No ID column with doubles
> 3) ID column with vector
> 4) No ID column with vector
> When feeding a DataFrame input into the new MLContext API, the DataFrame type can be specified as a MatrixFormat.
> For usability, the DataFrame type should also be dynamically determined from the data structure and data itself.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)