This package implements nonparametric smoothing methods described in the book of Bowman & Azzalini (1997)

Missing data are allowed; they are simply removed, togeter with
the associated variates from the same case, if any.
Datasets of arbitrary size can be handled by the current version of
`sm.density`

, `sm.regression`

and `sm.ancova`

, using
binning operations.

The functions in the package use kernel methods to construct nonparametric estimates of density functions and regression curves in a variety of settings, and to perform some inferential operations.

Specifically, density estimates can be constructed for 1-, 2- and 3-dimensional data. Nonparametric regression for continuous data can be constructed with one or two covariates, and a variety of tests can be carried out. Several other data types can be handled, including survival data, time series, count and binomial data.

The main functions are `sm.density`

and `sm.regression`

; other
functions intended for direct access by the user are:
`h.select`

, `binning`

,
`sm.ancova`

, `sm.autoregression`

, `sm.binomial`

,
`sm.binomial.bootstrap`

, `sm.poisson`

, `sm.poisson.bootstrap`

,
`sm.options`

, `sm.rm`

, `sm.script`

, `sm.sphere`

,
`sm.survival`

, `sm.ts.pdf`

, `sm.variogram`

, `sm.pca`

. There are undocumented functions which are called by these.

The function `sm.script`

is used to run a set of examples (called
scripts) presented in the book quoted below. These scripts are
associated with the package but the package can be used independently of
them. The scripts are generally based on the functions of the package
`sm`

, but a few of them make used of the `gam`

package.

R version >= 2.1.0. Packages `gam`

and `akima`

are used by
some of the scripts launched via `sm.script`

, but they are not
used by the functions of this package.

This is version 2.1. The most recent version of the package can be obtained from either of the web pages: http://www.stats.gla.ac.uk/~adrian/sm, http://azzalini.stat.unipd.it/Book\_sm

The book by Bowman and Azzalini (1997) provides more detailed and
background information. Algorithmic aspects of the software are
discussed by Bowman & Azzalini (2003). Differences between the first
version of the package, described in the book, and the current one are
summarized in the file `history.txt`

which is distributed with
the package.

Important contributions to prototype versions of functions for some specific techniques included here were made by a succession of students; these include Stuart Young, Eileen Wright, Peter Foster, Angela Diblasi, Mitchum Bock and Adrian Hines. We are grateful for all these interactions. These preliminary version have been subsequently re-written for inclusion in the public release of the package, with the exception of the functions for three-dimensional density estimation, written by Stuart Young. We also thank Luca Scrucca who made useful comments and who has ported the software to XLispStat. We are particularly grateful to Brian Ripley for substantial help in the production of installation files, the creation of MS-Windows versions, initial porting of the software from S-Plus to R and for maintaining the package on CRAN for several years.

This package and its documentation are usable under the terms of the "GNU General Public License", a copy of which is distributed with the package.

Adrian Bowman (Dept Statistics, University of Glasgow, UK) and Adelchi Azzalini (Dept Statistical Sciences, University of Padua, Italy). Please send comments, error reports, etc. to the authors via the web pages mentioned above.

Bowman, A.W. and Azzalini, A. (1997).
*Applied Smoothing Techniques for Data Analysis: *
*the Kernel Approach with S-Plus Illustrations.*
Oxford University Press, Oxford.

Bowman, A.W. and Azzalini, A. (2003).
Computational aspects of nonparametric smoothing
with illustrations from the `sm`

library.
*Computational Statistics and Data Analysis*, **42**, 545–560.

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.