flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Till Rohrmann (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-1749) Add Boosting algorithm for ensemble learning to machine learning library
Date Fri, 22 Apr 2016 09:11:12 GMT

    [ https://issues.apache.org/jira/browse/FLINK-1749?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15253589#comment-15253589
] 

Till Rohrmann commented on FLINK-1749:
--------------------------------------

Hi [~dedrummond],

I think Narayana is no longer working on this. Thus, you could take this issue over. But you're
right that it may make more sense to first start working on the weak learners. For the multinomial
logistic regression I haven't seen any progress yet. So I would assume that you can take this
one over, too. For the decision trees there is a PR but the community didn't have time to
review it yet.

> Add Boosting algorithm for ensemble learning to machine learning library
> ------------------------------------------------------------------------
>
>                 Key: FLINK-1749
>                 URL: https://issues.apache.org/jira/browse/FLINK-1749
>             Project: Flink
>          Issue Type: New Feature
>          Components: Machine Learning Library
>            Reporter: Till Rohrmann
>              Labels: ML
>
> Boosting [1] can help to create strong learners from an ensemble of weak learners and
thus improving its performance. Widely used boosting algorithms are AdaBoost [2] and LogitBoost
[3]. The work of I. Palit and C. K. Reddy [4] investigates how boosting can be efficiently
realised in a distributed setting. 
> Resources:
> [1] [http://en.wikipedia.org/wiki/Boosting_%28machine_learning%29]
> [2] [http://en.wikipedia.org/wiki/AdaBoost]
> [3] [http://en.wikipedia.org/wiki/LogitBoost]
> [4] [http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6035709]



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

Mime
View raw message