Hi there,
I'm pondering the idea of contributing code to sklearn, more precisely
multi-armed bandits [1].
This kind of algorithms are related to collaborative filtering, recommender
systems and reinforcement learning.
I haven't found anything in sklearn on those, but I'm wondering if anyone
has started to work on one of these topics in a private branch ?
I've already coded the classical ones (context-free bandits) :
espilon-first, epsilon-greedy, UCB1, but I feel I need some guidance on how
to best integrate into sklearn.
More specifically:
- should I start a new dir like "sklearn/recommenders" or "sklearn/bandits"
or "sklearn/reinforcement" ?
- should I decide on my own to create an API like the one for classifiers ?
Thanks for any insights,
Eustache
[1] http://en.wikipedia.org/wiki/Multi-armed_bandit
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
