DCSmooth provides functions for fast nonparametric estimation of the trend surface of a functional or spatial model. Regression types included are kernel regression and local polynomial regression. An automatic bandwidth selector is included.
You can install the released version of DCSmooth from CRAN with:
install.packages("DCSmooth")
## Example
This basic example shows the two-dimensional estimation of a single gaussian peak
function (in `y.norm1`).
```r
library(DCSmooth)
# simulated data
y = y.norm1 + matrix(rnorm(101^2), nrow = 101, ncol = 101)
# smooth trend surface
y_dcs = dcs(y)
plot(y_dcs, plot_choice = 2)
# surface.dcs(y_dcs, plot_choice = 2)
For a detailed description, see the vignette of this package.
vignette("DCSmooth")