[jira] [Resolved] (SPARK-830) add DenseVector and SparseVector to mllib, and replace all Array[Double] with Vectors

["Xiangrui Meng (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SPARK-830?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Xiangrui Meng resolved SPARK-830.
---------------------------------

       Resolution: Fixed
    Fix Version/s: 1.0.0
         Assignee: Xiangrui Meng

MLlib v1.0 will support sparse data.

> add DenseVector and SparseVector to mllib, and replace all Array[Double] with Vectors
> -------------------------------------------------------------------------------------
>
>                 Key: SPARK-830
>                 URL: https://issues.apache.org/jira/browse/SPARK-830
>             Project: Spark
>          Issue Type: Improvement
>          Components: Spark Core
>    Affects Versions: 0.8.0
>            Reporter: Jason Day
>            Assignee: Xiangrui Meng
>             Fix For: 1.0.0
>
>
> currently machine learning models in mllib package use raw Array[Double] directly which is not portable and elegant.
> Replacing arrays with vectors can provide the following benefits:
> 1. Higher Performance. When the data are dense vectors, using array is fine, but when the data is sparse, using SparseVector can gain higher performance
> 2. Higher abstraction. Vectors can provide higher abstractions, which are elegant and intuitive, while Array[Double] is verbose.



--
This message was sent by Atlassian JIRA
(v6.2#6252)