A list of topics for a Google summer of code (GSOC) 2011

Online learning

Mentor : O. Grisel

Goal : Devise an intuitive yet efficient API dedicated to the incremental fitting of some scikit-learn estimators (on an infinite stream of samples for instance).

See this thread on the mailing list for a discussion of such an API. Design decision will be taken by implementing / adapting three concrete models:

• text feature extraction
• online clustering with sequential k-means
• generalized linear model fitting with Stochastic Gradient Descent (both for regression and classification)

Boosting

Mentor : Satra

Manifold learning

Mentor : Fabian Pedregosa

Dictionary Learning a.k.a. Sparse Coding

Mentor : Gael Varoquaux, Alex Gramfort, Alex Passos

Some useful ressources with compatible License:

• Sparse PCA gist

• Another gist by A. Passos

• NMF + Hoyer method in milk

Vlad candidate?

Random forest

Mentor : Satra

(there is already a preliminary implementation in my fork) i would combine this with boosting/bagging

Locality Sensitive Hashing

Mentor : Mathieu Blondel?

Command line interface

Mentor : ?

Interaction with mldata.org

Mentor : ?