Orange Forum • View topic - Regarding GSoc Project "Support for parallel computation"

Regarding GSoc Project "Support for parallel computation"

General discussions about Orange and with Orange connected things (data mining, machine learning, bioinformatics...).

Regarding GSoc Project "Support for parallel computation"

Postby vijaym123 » Mon Mar 19, 2012 15:25

Hello,

I am Vijay Mahantesh SM from India. I am an open source enthusiast and a big fan of computational mathematics. I came across the idea list of Orange and was fascinated to find projects of my passion.I have a good exposure in Python along with knowledge of certain subjects like linear algebra, machine learning, data mining .

I had previously worked on research projects in Complex Networks "Prediction Of Arrival Of Nodes In A Scale Free Network" http://www.arxiv.org/abs/1111.4886 and "A Navigation Algorithm Inspired by Human Navigation" link http://www.arxiv.org/abs/1111.4898. These project involves a good amount of machine learning concepts. Have implemented these projects in python along with supporting libraries such as NetworkX, Matplotlib, Numpy,Scipy.

I have understood the gist of the project along with Cross-Validation concept. Wanted to know some particular details with the project,Like which distributed computing package ( eg: Hadoop ) is to be used ? Some homework, which i need to be prepared with for getting along with this project ?

Awaiting reply ! :)

Regards,
Vijay Mahantesh SM
6th Sem,
PESIT,
Bangalore,
India.

Re: Regarding GSoc Project "Support for parallel computation

Postby Anze » Tue Mar 20, 2012 14:59

Orange does not support any kind of parallel computation, so its it entirely up to you to decide, which framework for parallelization will be used. You should be familiar with what can and cannot be done with multiprocessing module (http://docs.python.org/library/multipro ... processing) and compare other frameworks (if you choose them) against it.

You should be able to identify at least few parts of orange that could benefit from the support for parallel computation by the time you submit your final proposal. The proposal should not be too general. You need to show us, that you understand what orange does and what kind of parallelization it needs.


Return to General Discussions



cron