BARS code is now available as R packages for Linux and Windows:
package for Linux
package for Windows
Please note: in the windows version the R wrapper line
cmdpath <- file.path(pkgpaht,"exec/barsN.out")
should be cmdpath <- file.path(pkgpaht,"exec/barsN.exe")
[we haven't changed this yet]
Please address questions to Jeff Liebner (email@example.com).
Help with installation can be found at The Comprehensive R Archive Network.
Our paper for the Journal of Statistical Software provides a
detailed description of
the BARS code.
The Poisson version of BARS is also available for Matlab:
BARS for Matlab
Questions about the Matlab version
should be addressed to Ryan Kelly (firstname.lastname@example.org).
Papers applying (or modifying) BARS include:
Behseta, S., Wallstrom, G.L., and Kass, R.E. (2005, Biometrika)
Behseta, S. and Kass, R.E. (2005, Statistics in Medicine)
Kaufman, C.G., Ventura, V., and Kass, R.E. (2005, Statistics in
Wallstrom, G., Liebner, J., and Kass, R.E. (2005)
Kass, R.E., Ventura, V. and Cai, C. (2003, Network)
Wallstrom, G.L, Kass, R.E., Miller, A., Cohn, J.F., and Fox,
N.A. (2002, Case Studies in Bayesian Statistics)
Dimatteo, I., Genovese, C.R., and Kass, R.E. (2001, Biometrika)
These may be obtained from my recent publications page.
BARS (Bayesian Adaptive Regression Splines) solves
the generalized nonparametric regression (curve-fitting) problem
by assuming the function may be approximated by a spline.
Here, for example, the data may be binary, or counts, and the explanatory
variable may be time. The special cases in which
the data are continuous pose the usual curve-fitting problem,
ordinarily solved by some variation on least-squares.
A substantial literature has
demonstrated the power of spline-based generalized curve-fitting.
See Hansen and Kooperberg (2002, Statist. Science) for a review.
The difficult part of the problem is to allow aspects of the spline to
vary (adaptively to the data) across the domain of .
DiMatteo, Genovese, and Kass (2001, Biometrika) proposed BARS and
contributed an initial implementation and study of the method.
- prior information about the
number of knots is formulated as a prior probability distribution (e.g.,
uniform on the integers from 1 to a maximal value );
- reversible-jump MCMC on the knot sets;
- Laplace's method, to integrate the spline coefficients;
- continuous proposals for knot locations, which attempt to place
new knots near existing knots; and,
- in existing implementations,
``unit-information priors'' on the spline coefficients, given the knot
set. (This is not essential to the method.)
- BARS computes
- a posterior distribution on the knot sets, and
- a posterior
distribution on any set of function values .
DiMatteo et al. compared BARS to two recently successful
methods of solving the usual curve-fitting problem.
- Denison, Mallick, and Smith (DMS; JRSSB, 1998): a quasi-Bayesian
approach, which they found provided smaller MSE than wavelet
fits in examples from Donoho and Johnstone (1995).
- Zhou and Shen (SARS; JASA, 2001): an optimization method,
which they said
``performed at least as well as the spline competitors in all ...
examples, and significantly better in some,'' and again provided smaller MSE
than wavelet fits in examples from Donoho and Johnstone (1995).
A typical data set simulated from a true curve, together with fits
each of DMS, SARS, and BARS are shown in the following figure.
The fits are all a bit more wiggly than the true curve, but BARS
provides a smoother fit while still capturing the sudden
jump. Mean-squared errors in several examples were much smaller for
BARS than for DMS or SARS.
The next figure shows
a BARS Poisson regression fit (thick curve) to neuronal data,
providing the kind of smoothing we believe to be desirable; also shown is
a Gaussian kernel density (Gaussian filter) estimate (thin curve).
Taken from Kass, Ventura, Cai (2003, NETWORK:
Computation in Neural Systems).