Priorities

(1) Refactor R objects and routines. Figure out an appropriate naming
    convention for these.
    -->> DONE

(2) Implement and wrap resampled ensemble feature selection. Implement and wrap function to return the mRMR score for each solution.
    -->> DONE

(3) Provide documentation for all public methods.
    (a) Man pages
    (b) Vignette
    -->> DONE

(4) Add a new dataset with breast cancer (BC) prognosis so we show 2 applications of mRMRe in the application note. So 2 datasets will be used:
    1. Irinotecan sensitivity prediction: train in CGP and validate in CCLE
    2. BC prognosis: train in VDX ER-/HER2- and validate on TRANSBIG ER-/HER2-
    -->> DONE (see Dropbox/mRMRe_appnote/revision/analysis/data)

(4) With these 2 datasets, perform stability-diversity and GO analysis to compare classic, resampled and parallel mRMR feature selections
    (a) Resolve the feature # vs solution length debacle
    (b) Find the best way to report GO results
    -->> DONE

(5) Update the article.
    (a) Regenerate (and document) experiments discussed (with the resampled
        ensemble feature selection and use only 100 solutions, show the distribution of mRMR scores)
    (b) Add mRMR formula
    (c) Clearly state the 2 ensemble appraoches we have implemented
        resampled: multiple datasets / 1 solution each
        parallel: 1 dataset / multiple solutions
    (d) Add stability-diversity analysis and update GO analysis results
    -->> DONE





Would be awesome to have

(*) Try to solve compiler issues with the package. I propose that we avoid
    hard-coding any hacks and that we instead provide the user with tips in
    the Installation section of the vignette (I have thought of several hacks
    that could mitigate the issues we've been experiencing). 

(*) A Nobel prize for our awesome work (it could be in the luxury section).

Luxury (move to the section above if deemed necessary)

( ) Sit down and think really hard about how sample weighing and stratification
    is implemented for measuring association between features.

( ) Improve, automate and integrate the workflow.

  ( ) Bug-tracking system. (JIRA?)

  ( ) Unit, integrative and systems testing

( ) Document internal C++ code.
