Local Regression and Likelihood

LOCFIT

Local Regression and Likelihood

I am moving to Case Western Reserve University in August 2002. Updated Locfit pages may be found on the CWRU statistics department web site sometime in the future.

LOCFIT is a software system for fitting curves and surfaces to data, using the local regression and likelihood methods. The code is mostly written in C; an interface is provided that enables LOCFIT to be used as an S-Plus or R library.
For a quick tour of LOCFIT's main capabilities, you may like this introductory paper.

Download LOCFIT

For S and S-Plus users

Windows
- S-Plus 6.0 self-extracting zip.
  Last updated January 10, 2002.
- S-Plus 4.0,4.5,2000 self-extracting zip.
  Last updated February 25, 2001.
- S-Plus 3.3 self-extracting zip.
  Last updated November 10, 2000.
- Installation
Unix, Linux
- Source Code for S-Plus 5, 6. [Installation]
  Last updated August 2, 2001.
- Compiled distributions for S-Plus 5, 6.
  Last updated August 2, 2001.
- Source Code for S-Plus 3. [Installation]
  Last updated June 14, 1999. This is an old archive -- I haven't maintained the installation files for S-Plus 3 in the newer archive, but the code should still be compatible.

For R users

Source Code [Installation]
Last updated August 3, 2001.

Note: Some binary distributions of R have LOCFIT preinstalled. Type library() at the R prompt to get a list of your currently installed libraries.

As a stand-alone program, C-Locfit

Due to time constraints, I do very little maintainence and development of these versions. I encourage using Locfit as an S-Plus or R library instead. R is free.

Source Code for unix users. [Installation]
Last updated August 2, 2001.
Windows 95/98/2000/NT compiled code. [Installation]
Last updated August 3, 2001.

Documentation

R and S Help Pages

On-line help pages for functions in the R and S versions of Locfit.
Download all help files for local installation here (tar,gzipped).
Last updated November 10, 2000.

Local Regression and Likelihood

Book published by Springer (1999). This reference describes the local regression and likelihood procedures and associated methods, and contains numerous examples using LOCFIT.
Additions, clarifications and corrections to the book can be found here. [PDF version]
Last updated August 3, 2001.
S functions for the examples in the book. See the README file in the archive for instructions.