Performance improvement by compiling using Cython

[pramitchoudhary]

Reference blog: https://www.ibm.com/developerworks/community/blogs/jfp/entry/A_Comparison_Of_C_Julia_Python_Numba_Cython_Scipy_and_BLAS_on_LU_Factorization?lang=en

It will interesting to see if one can improve performance of pdp and variable importance by specifying static types and compiling the same using Cython.

First step to start in this direction could be optimized the dataframes usage, e.g. https://pandas.pydata.org/pandas-docs/stable/enhancingperf.html

[aikramer2]

@pramitchoudhary there arent really any opportunities for this at the moment. The things we could port to cython make up a very small part of the computational footprint.