R-package-pointers.Rmd

# R Packages for Mixed(-Effects) Models

## The absolute core:

### [`nlme`](https://CRAN.R-project.org/package=nlme)
### [`lme4`](https://CRAN.R-project.org/package=lme4)

## Other Packages with Mixed Effect Models as main purpose (and recommendable)

- [`lmerTest`](https://CRAN.R-project.org/package=lmerTest)
- [`glmmTMB`](https://CRAN.R-project.org/package=glmmTMB)

- [`pbkrtest`](https://CRAN.R-project.org/package=pbkrtest), see also the [`pbkrtest` page](http://people.math.aau.dk/~sorenh/software/pbkrtest/)    and notably the `demo.R` and resulting [`demo.html`](http://people.math.aau.dk/~sorenh/software/pbkrtest/doc/demo.html) example.

## [CRAN Task Views](https://cran.r-project.org/web/views)

There is none, with this topic
Is there no good one ?

- `nlme` is in
	`ChemPhys`,
	`Econometrics`,
	`Environmetrics`,
	`Finance`,
	`OfficialStatistics`,
	`Psychometrics`,
	`SocialSciences`,
	`Spatial`,
	`SpatioTemporal`,

	whereas

- `lme4` is in
	`Bayesian`,
	`Econometrics`,
	`Environmetrics`,
	`OfficialStatistics`,
	`Psychometrics`,
	`SocialSciences`,
	`SpatioTemporal`,

- with the intersection

| Task View          | URL                                           | CRAN packages | comments |
|--------------------|-----------------------------------------------|---------------|----------|
| `Econometrics`     | https://CRAN.R-project.org/view=Econometrics | nlme, lme4 |   |
| `Environmetrics`   | https://CRAN.R-project.org/view=Environmetrics | nlme, lme4 | Modelling species responses and other data   |
|`OfficialStatistics`| https://CRAN.R-project.org/view=OfficialStatistics | nlme, lme4, rsae | (robust) Small Area Estimation  |
| `Psychometrics`    | https://CRAN.R-project.org/view=Psychometrics | nlme, lme4, ordinal, MCMCglmm, irtrees | Item Response Theory (ITR), Rasch Mod.  |
| `SocialSciences`   | https://CRAN.R-project.org/view=SocialSciences | nlme, lme4, lmeSplines, lmm, MCMCglmm | Other Regression Methods |
| `SpatioTemporal`   | https://CRAN.R-project.org/view=SpatioTemporal | nlme, lme4 | (spatio-temporal data)  |

## R (CRAN) packages with `LMM` in their names:
```{r, lmm-pkgs, cache=TRUE}
ip <- installed.packages() # ~ 13'000 for Martin Maechler
str(ip)
pkgs <- rownames(ip)
str(LMMs <- grep("lmm", pkgs, ignore.case=TRUE, value=TRUE))
pdLMM <- lapply(LMMs, packageDescription)
titLMM <-
    ttLMM <- gsub("\n", " ",
                  sapply(setNames(,LMMs), packageDescription, fields = "Title"))
longT <- nchar(ttLMM) > 80
ttLMM[longT] <- paste(substr(ttLMM[longT], 1, 77), "...")
cbind(ttLMM) # 80% are GLMM  amazing
```
<!-- When run by  `R RHOME`/bin/Rscript --no-restore --no-save -e 'knitr::knit("R-package-pointers.Rmd")'
 this fails with (R <-> R-devel) problem (and cache ?):
 Error in readRDS(file) :
  cannot read unreleased workspace version 3 written
  by experimental R 3.5.0 -->


## Closely related packages

- [`ordinal`](https://CRAN.R-project.org/package=ordinal), fn `clmm()` for *cumulative* LMMs.
- [`lmeSplines`](https://CRAN.R-project.org/package=lmeSplines), based on pkg `nlme`, allows splines for linear and non-linear MM.

- [`mgcv`](https://CRAN.R-project.org/package=mgcv), *Recommended* pkg, has `gamm()`, based on pkg `nlme`.  Alternatively, (and not possible inside *rec.* package `mgcv`):
- [`gamm4`](https://CRAN.R-project.org/package=gamm4), based on pkg `lme4`, has fn. `gamm4()` based on `lmer()` etc.


#### NB. "knit me" to markdown --> `*.md` simply by

	od <- setwd("~/R/D/GH/MEMo"); knitr::knit("R-package-pointers.Rmd") ; setwd(od)