mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Dmitriy Lyubimov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAHOUT-1490) Data frame R-like bindings
Date Wed, 21 May 2014 00:19:39 GMT

    [ https://issues.apache.org/jira/browse/MAHOUT-1490?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14004149#comment-14004149
] 

Dmitriy Lyubimov commented on MAHOUT-1490:
------------------------------------------

i did not say it was bad. I said the same things, it is good for certain algorithms. Actually,
a lot of algorithms, of iterative nature.

I am just thinking how to expose cost to algorithm layer so it doesn't do naive things. The
programming model, it is all about it. imagine an algorithm that does something like gaussian
elimnation or givens QR. Obviously compression doesn't help here since inflate/deflate cycle
will cost more than any benefits of compressed reads, it would seem it would be faster with
just uncompressed vectors.

Fortunately we don't have to care about delayed updates since we are doing 100% in-core local
operation here. 

> Data frame R-like bindings
> --------------------------
>
>                 Key: MAHOUT-1490
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-1490
>             Project: Mahout
>          Issue Type: New Feature
>            Reporter: Saikat Kanjilal
>            Assignee: Dmitriy Lyubimov
>             Fix For: 1.0
>
>   Original Estimate: 20h
>  Remaining Estimate: 20h
>
> Create Data frame R-like bindings for spark



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

Mime
View raw message