# sm.rm: Nonparametric analysis of repeated measurements data In sm: Smoothing Methods for Nonparametric Regression and Density Estimation

## Description

This function estimates nonparametrically the mean profile from a matrix `y` which is assumed to contain repeated measurements (i.e. longitudinal data) from a set of individuals.

## Usage

 ```1 2``` ```sm.rm(Time, y, minh = 0.1, maxh = 2, optimize = FALSE, rice.display = FALSE, ...) ```

## Arguments

 `y` matrix containing the values of the response variable, with rows associated to individuals and columns associated to observation times. `Time` a vector containing the observation times of the response variable, assumed to be the same for all individuals of matrix `y`. If `Time` is not given, this is assumed to be `1:ncol(y)`. `minh` the mimimum value of the interval where the optimal value of the smoothing parameter is seached according to the modified Rice criterion. See reference below for details. `maxh` the maximum value of the above interval. `optimize` Logical value, default is `optimize=FALSE`. If `optimize=TRUE`, then a full optimization is performed after searching the interval `(minh,maxh)` using the optimizer `optim`. `rice.display` If this set to `TRUE` (default is `FALSE`), a plot is produced of the curve representing the modified Rice criterion for bandwidth selection. See reference below for details. `...` other optional parameters are passed to the `sm.options` function, through a mechanism which limits their effect only to this call of the function; those relevant for this function are the following: add logical value, default is `add=FALSE`. If `add=TRUE` and display is not set to `"none"`, then graphical output added to the existing plot, rather than starting a new one. display character value controlling the amount of graphical output of the estimated regression curve. It has the same meaning as in `sm.regression`. Default value is `display="lines"`. ngrid the number of divisions of the above interval to be considered. Default: `ngrid=20`. poly.index overall degree of locally-fitted polynomial, as used by `sm.regression`. Default: `ngrid=1`.

## Details

see Section 7.4 of the reference below.

## Value

a list containing the returned value produced by `sm.regression` when smoothing the mean response value at each given observation time, with an extra component `\$aux` added to the list. This additional component is a list itself containing the mean value at each observation time, the residual variance of the residuals from the estimated regression curve, the autocorrelation function of the residuals, and the value `h` of the chosen smoothing parameter.

## Side Effects

if the parameter display is not set to `"none"`, a plot of the estimated regression curve is produced; other aspects are controlled by the optional parameters (`...`). If `rice.display=TRUE`, a plot of the modified Rice criterion is shown.

## References

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

## See Also

`sm.regression`, `sm.regression.autocor`, `optim`

## Examples

 ```1 2 3 4 5 6 7 8``` ```sm.rm(y=as.matrix(citrate), display.rice=TRUE) # with(dogs, { Time <- seq(1,13,by=2) gr1 <- as.matrix(dogs[dogs\$Group==1,2:8]) plot(c(1,13), c(3,6),xlab="time", ylab="potassium", type="n") sm1 <- sm.rm(Time, gr1, display="se", add=TRUE) }) ```

sm documentation built on May 1, 2019, 8:06 p.m.