ksmooth
Kernel Regression Smoother
Description
Performs scatterplot smoothing using kernel estimates.
Usage
ksmooth(x, y, kernel = c("box", "normal"), bandwidth = 0.5,
range.x = range(x), n.points = max(100, length(x)), x.points)
Arguments
| x |
a vector of x data.
Missing values (NAs) are not accepted.
|
| y |
a vector of y data.
This must be same length as x, and
missing values (NAs) are not accepted.
|
| kernel |
a character string that determines the smoothing kernel.
kernel can be one of the following.
|
| "box" | a rectangular box (the default). | |
| "normal" | the gaussian density function.
|
|
| bandwidth |
the kernel bandwidth smoothing parameter.
All kernels are scaled so the upper and lower quartiles of the
kernel (viewed as a probability density) are +/- 0.25.
Larger values of bandwidth make smoother estimates;
smaller values of bandwidth make less smooth estimates.
|
| range.x |
a vector containing the minimum and maximum values of x at which to
compute the estimate. If not specified,
this argument is set to range(x).
|
| n.points |
the number of points to smooth in the interval range.x.
n.points is set to the length of x.point
if x.points is supplied.
|
| x.points |
a vector specifying where the kernel estimate is computed.
If this argument is not specified,
it is set to seq.int(range.x[1], range.x[2], length=n.points).
|
Value
returns a list containing the following components:
| x |
a vector of sorted x values at which the kernel estimate was computed.
|
| y |
a vector of smoothed estimates for the regression at the corresponding x.
|
References
Silverman, B. W. 1986. Density Estimation for Statistics and Data Analysis. London, UK: Chapman and Hall.
Watson, G. S. 1966. Smooth regression analysis. Sankha, Ser. A. Volume 26. 359-378.
See Also
Examples
# Make up some data:
set.seed(50)
x <- rnorm(100)
eps <- rnorm(100, 0, .1)
y <- sin(x) + eps
# A kernel regression estimate for y on x.
ksmxy <- ksmooth(x, y, kern="normal", band=.5, n=50)