Description

mlr is an R package that integrates almost all machine learning packages in R. RMVA is a set of plugins for TMVA based on ROOTR that allows the use of R in TMVA. The idea is to update the library RMVA to use mlr.

Task ideas

• Write BaseMethod class for mlr package.
• Implement a system to check if the required machine learning method has the R package installed.
• Implement a system to check if the required machine learning method from R has support for multiclass classification, regression or weights in the data.
• Write methods to support classification and regression.
• Write a clear documentation with a table of options for every method.

Expected results:

• working implementation with tests and documentation
• A machine learning example using different methods for classification and regression.
• Compare C50/RSVM and RSNNS using the current RMVA implementation and with the new MLR implementation.

Requirements: Advanced skills in C/C++, experience with R and machine learning.

Mentors:

• Omar Zapata
• Lorenzo Moneta
• Diego Restrepo

Links:

• http://root.cern
• http://mlr-org.github.io
• http://oproject.org/RMVA
• http://oproject.org/ROOT+R

Corresponding Project

• TMVA

Participating Organizations

• UdeA
• CERN

---