# sm.autoregression: Nonparametric estimation of the autoregression function In sm: Smoothing Methods for Nonparametric Regression and Density Estimation

## Description

This function estimates nonparametrically the autoregression function (conditional mean given the past values) of a time series `x`, assumed to be stationary.

## Usage

 ```1 2``` ```sm.autoregression(x, h = hnorm(x), d = 1, maxlag = d, lags, se = FALSE, ask = TRUE) ```

## Arguments

 `x` vector containing the time series values. `h` the bandwidth used for kernel smoothing. `d` number of past observations used for conditioning; it must be 1 (default value) or 2. `maxlag` maximum of the lagged values to be considered (default value is `d`). `lags` if `d==1`, this is a vector containing the lags considered for conditioning; if `d==2`, this is a matrix with two columns, whose rows contains pair of values considered for conditioning. `se` if `se==T`, pointwise confidence bands are computed of approximate level 95%. `ask` if `ask==TRUE`, the program pauses after each plot until is pressed.

## Details

see Section 7.3 of the reference below.

## Value

a list with the outcome of the final estimation (corresponding to the last value or pairs of values of lags), as returned by `sm.regression`.

## Side Effects

graphical output is produced on the current device.

## 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.

`sm.regression`, `sm.ts.pdf`

## Examples

 ```1 2``` ```sm.autoregression(log(lynx), maxlag=3, se=TRUE) sm.autoregression(log(lynx), lags=cbind(2:3,4:5)) ```

### Example output     ```Package 'sm', version 2.2-5.6: type help(sm) for summary information
Warning message:
no DISPLAY variable so Tk is not available
```

sm documentation built on Sept. 13, 2021, 5:07 p.m.