DenseFeatureGroups

C++ implementation of the article http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.155.896

Usage

mkdir build
cd build
cmake ..
make
./kaspar

You will see [0, 0, 0], [0, 1, 2]

Purposes

1. You can provide your own smoothing kernels
2. Use function ObtainFeatureGroups
   1. kernel to use (K : Eigen::VectorXd -> double)
   2. data matrix, each row is a unique feature of all data points (Eigen::MatrixXd)
   3. h - bandwidth for the algorithm
3. Link your code with generated library libDGF