[
https://issues.apache.org/jira/browse/SPARK-4638?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14300012#comment-14300012
]
Mandar Chandorkar commented on SPARK-4638:
------------------------------------------
Hello all, I have been working on an implementation of Kernels/Kernel Matrices
as a part of my Masters thesis at KU Leuven.
I have implemented RBF & Polynomial Kernels for SVMs. I have drawn up a
hierarchy of classes/interfaces which can be extended to implement other
Kernels as well. I have made a new module within the mllib code. This is my
first attempt at contributing to spark (and open source) so my code might be
messy. To summarize I have worked on the following.
1) Class hierarchy for SVM Kernels, with unit tests.
2) Entropy based subset selection for low rank approximation of Large Kernel
Matrices.
3) Kernels for density estimation, with 'plug in' based optimum bandwidth
selection.
I have the code on a local branch, I can push it to my fork and do a pull
request, but before that is there anything else I should do (apart from
checking the code style and all that)? I can also make a short design document
describing the class hierarchies and how they are connected.
Thank you
> Spark's MLlib SVM classification to include Kernels like Gaussian / (RBF) to
> find non linear boundaries
> -------------------------------------------------------------------------------------------------------
>
> Key: SPARK-4638
> URL: https://issues.apache.org/jira/browse/SPARK-4638
> Project: Spark
> Issue Type: New Feature
> Components: MLlib
> Reporter: madankumar s
> Labels: Gaussian, Kernels, SVM
>
> SPARK MLlib Classification Module:
> Add Kernel functionalities to SVM Classifier to find non linear patterns
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
