Raw File
Tip revision: e6bcdf04d750769ae65b0b3951329e77ef9876bd authored by Kristy Robledo on 15 May 2023, 22:50:02 UTC
version 2.0
Tip revision: e6bcdf0

<!-- is generated from README.Rmd. Please edit that file -->

# VarReg

<!-- badges: start -->

<!-- badges: end -->

<!-- Please don't remove this: Grab your social icons from -->
<!-- display the social media buttons in your README -->
<!--[![alt text][1.1]][1] -->
<!-- links to social media icons -->
<!-- no need to change these -->
<!-- icons with padding -->
<!-- icons without padding -->
<!-- links to your social media accounts -->
<!-- update these accordingly -->
<!-- Please don't remove this: Grab your social icons from -->

The goal of VarReg is to provide methods for fitting semi-parametric
mean and variance models, with normal or censored data. This has also
been extended to allow a regression in the location, scale and shape
parameters. This algorithm is based upon an EM (Expectation
Maximisation) algorithm, so is more stable than other similar methods
like GAMLSS.

## :raising_hand: Author

Kristy Robledo <>

NHMRC Clinical Trials Centre, University of Sydney

## :arrow_double_down: Installation

You can install the released version of VarReg from
[CRAN]( with:

``` r

And the development version from [GitHub]( with:

``` r
# install.packages("devtools")

## :book: Examples

This is a basic example to read in the mcycle dataset and perform a
linear model in the mean and the variance:

``` r
#> Welcome to the 'VarReg' package to perform semi-parametric regression
## read in dataset
## run a model with linear mean and linear variance:
linmodel<-semiVarReg(mcycle$accel, mcycle$times, meanmodel="linear", varmodel="linear",

Now we can plot the model:

``` r

<img src="man/figures/README-plotlinear-1.png" width="100%" />

``` r

##can also add CI
plotVarReg(linmodel, ci=TRUE, ci.type = "im")
#> [1] "CI=true, type=information matrix"

<img src="man/figures/README-plotlinear-2.png" width="100%" />

Or we can look at the results:

``` r
#> [1] -683.5092

#>    Intercept mcycle$times 
#>    -53.69517      1.11797

#>    Intercept mcycle$times 
#>   3824.07225    -66.39011

We can also run a model with semi-parametric mean (4 internal knots) and
semi-parametric variance (2 knots):

``` r
semimodel<-semiVarReg(mcycle$accel, mcycle$times, meanmodel="semi", varmodel="semi",
knots.m=4, knots.v=2, maxit=10000)

<img src="man/figures/README-semimodel-1.png" width="100%" />

``` r

## run a model with semi-parametric mean (4 internal knots) and semi-parametric monotonic
## variance (2 knots):
## not run
##semimodel_inc<-semiVarReg(mcycle$accel, mcycle$times, meanmodel="semi", varmodel="semi",
##knots.m=4, knots.v=2, mono.var="inc")

Lastly, we can fit a model with a model in the location, scale and
shape. Im not going to run this, just show the code, as it takes a while
to run on my laptop!

``` r
 ##  LSS model followed by the basic plot command
#lssmodel<-lssVarReg(mcycle$accel, mcycle$times,  locationmodel="linear", scale2model="linear", shapemodel="constant", maxit=10000)
#plotlssVarReg(lssmodel, xlab="Time in seconds", ylab="Acceleration")

back to top