Targeted Learning of a NP Importance of a Continuous Exposure

Targeted minimum loss estimation (TMLE) of a non-parametric variable importance measure of a continuous exposure 'X' on an outcome 'Y', taking baseline covariates 'W' into account.


Package: tmle.npvi

Version: 0.10.0 [2015-05-22] o Added CITATION file. o Added reference to Bioinformatics Application Note. o Enhanced the parsimonious simulation of 'X' given 'W' thanks to a discussion with Emily Chang (UCSF). Side effect: package does not need to import 'sgeostat::in.chull' and 'sgeostat::in.polygon' anymore. o Passes R CMD check. o The function 'tmle.npvi' now also handles data frames.

Version: 0.9.3 [2015-02-05] o Shortened package title as per the CRAN policies.

Version: 0.9.2 [2015-02-05] o Updated NAMESPACE to make use of S3 method registration. o Dropped support of 'DSA' SuperLearning libraries, as DSA is not a 'mainstream package'.

Version: 0.9.1 [2014-12-12] o Added 'getPValue' for testing \eqn{Psi(P_0)=Phi(P_0)}'' or\eqn{Psi(P_0)=0}''.

Version: 0.9.0 [2014-11-14] o Faster version, which handles much larger data sets:

  • Now using sparse matrices to speed up the computations of the tabulated versions of the features of interest. Based on new parameter 'nMax' in 'getSimulationScheme'.
  • Faster version of 'simulateParsimiouslyYgivenXW' and 'simulateParsimiouslyXgivenW'
  • Added parameter 'nMax' to speed up 'getSimulationScheme'. o Package does not DEPEND on 'MASS' and 'sgeostat' anymore, but IMPORTS functions from them. o Added tcga12brca data set. o Updated DESCRIPTION as suggested by Henrik Bengtsson. o Enhanced scripts for TCGA data analysis.

Version: 0.8.1 [2014-02-08] o Now using lazy-loading of learning libraries instead of assigning objects to the global environment.

Version: 0.8.0 [2014-02-07] o Passes R CMD check --as-cran


Version: 0.1.0 [2010-07-13] o Passes R CMD check. o Created.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.10.0 by Pierre Neuvial, 6 years ago

Browse source code at

Authors: Antoine Chambaz , Pierre Neuvial

Documentation:   PDF Manual  

GPL license

Imports R.methodsS3, R.oo, MASS, Matrix, geometry

Depends on R.utils

Suggests SuperLearner, e1071, randomForest, polspline, gam, knitr

See at CRAN