[R-pkgs] New package `np' - nonparametric kernel smoothing methods for mixed datatypes

Jeffrey Racine racinej at mcmaster.ca
Fri Nov 24 20:26:04 CET 2006

Dear R users,

A new package titled `np' is now available from CRAN.

The package implements recently developed kernel methods that seamlessly
handle the mix of continuous, unordered, and ordered factor datatypes
often found in applied settings. 

The package also allows users to create their own
nonparametric/semiparametric routines using high-level function calls
(via the function npksum()) rather than writing their own C or Fortran
code. Much of the code underlying the package is written in C including
the function npksum().

Currently, a range of methods can be found in the package including

- multivariate nonparametric unconditional and conditional density

- multivariate nonparametric conditional mean and gradient estimation
(local constant and local linear)

- multivariate nonparametric conditional distribution, quantile and
gradient estimation 

- nonparametric model specification tests for testing correctness of
parametric regression models

- nonparametric significance tests for nonparametric regression

- semiparametric multivariate regression methods (partially linear,
index models, average derivative estimation, varying/smooth coefficient

A function npplot() is implemented that allows users to visualize the
resulting objects.

A variety of methods for computing standard errors and error bounds are
implemented including asymptotic and resampling-based methods (the
latter employing the `boot' package which is required).

A range of examples can be found in the examples section of the
accompanying help files.

We caution potential users that multivariate data-driven bandwidth
selection methods can be numerically intensive. For this reason we are
now using some of the functionality contained in the Rmpi package to
develop an MPI-aware version of the np package that we have tentatively
titled the `npRmpi' package.

The np package implements a number of methods found in the newly
released publication by Li and Racine (2007) titled Nonparametric
Econometrics: Theory and Practice, Princeton University Press.

See press.princeton.edu/titles/8355.html for further details.

Best regards, Jeff.

Professor J. S. Racine         Phone:  (905) 525 9140 x 23825
Department of Economics        FAX:    (905) 521-8232
McMaster University            e-mail: racinej at mcmaster.ca
1280 Main St. W.,Hamilton,     URL:
Ontario, Canada. L8S 4M4

`The generation of random numbers is too important to be left to chance'

More information about the R-packages mailing list