densitas is a C++ library for density estimation of regression problems.

With densitas you can estimate the quantiles of the unknown probability
distribution around the true value for each event.

densitas implements the approach first proposed in:
Predicting Conditional Probability Distributions: A Connectionist Approach, 1995
Weigend, A. S., and A. N. Srivastava, International Journal of Neural Systems

For further insight have a look at densitas/density_estimator.hpp

A real world example can be found at test/real_world_with_liblinear.cpp

To build and run the tests the following libraries are required:
- liblinear (http://www.csie.ntu.edu.tw/~cjlin/liblinear)
- armadillo (http://arma.sourceforge.net)
- libunittest (http://libunittest.sourceforge.net)

densitas itself has no dependencies except for the standard library.
Supported compilers are: g++ (>=4.6), clang++ (>=3.2), icc (>=14)
 
densitas is being developed by Christian Blume. Contact Christian at
chr.blume@gmail.com for any questions or comments.