https://github.com/sempwn/gamdiffs

Estimation of confidence intervals for GAMs
https://github.com/sempwn/gamdiffs

r r-package rstats

Last synced: about 1 month ago
JSON representation

Estimation of confidence intervals for GAMs

• Host: GitHub
• URL: https://github.com/sempwn/gamdiffs
• Owner: sempwn
• License: other
• Created: 2022-06-01T23:51:40.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-11-18T22:04:46.000Z (about 2 years ago)
• Last Synced: 2024-10-12T01:23:02.018Z (3 months ago)
• Topics: r, r-package, rstats
• Language: R
• Homepage: https://sempwn.github.io/gamdiffs/
• Size: 126 KB
• Stars: 0
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.Rmd
  • Contributing: .github/CONTRIBUTING.md
  • License: LICENSE
  • Code of conduct: .github/CODE_OF_CONDUCT.md

Awesome Lists containing this project

README

        
---

output: github_document

---

```{r, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>",

  fig.path = "man/figures/README-",

  out.width = "100%"

)

options(scipen = 1, digits = 2)

```

# gamdiffs

[![Lifecycle: experimental](https://img.shields.io/badge/lifecycle-experimental-orange.svg)](https://lifecycle.r-lib.org/articles/stages.html#experimental)

[![CRAN status](https://www.r-pkg.org/badges/version/gamdiffs)](https://CRAN.R-project.org/package=gamdiffs)

[![R-CMD-check](https://github.com/sempwn/gamdiffs/workflows/R-CMD-check/badge.svg)](https://github.com/sempwn/gamdiffs/actions)

The goal of gamdiffs is to provide a set of convenience functions for the estimation

and associated standard errors / confidence intervals for a range of Generalized

Additive Models. The particular motivation is to provide tools for estimating the 

difference in outcome under certain counterfactual scenarios for models relevant

to population public health.

Current implemented confidence interval estimates are

* delta method

* posterior sampling

* Efron bootstrap sampling

## Installation

You can install the released version of gamdiffs from [CRAN](https://CRAN.R-project.org) with:

``` r

install.packages("gamdiffs")

```

And the development version from [GitHub](https://github.com/) with:

``` r

# install.packages("devtools")

devtools::install_github("sempwn/gamdiffs")

```

## Example

This basic example generates some data and fit a thin-plate spline to a time

variable (`x`) with outcome `y` which is Poisson-distributed. The sum of the outcome

between time 0 to 20 is then compared to the outcome between time 21 to time 40 

and the resulting expected difference with associated confidence intervals is 

provided

```{r example}

library(gamdiffs)

library(mgcv)

## basic usage

res <- gamdiffs:::create_random_data()

m <- mgcv::gam(y ~ s(x), data = res, family = poisson)

baseline_data <- dplyr::tibble(x = 0:20)

counter_data <- dplyr::tibble(x = 21:40)

test_diffs <- calc_sum_counterfactual_gam(m, baseline_data,

  counter_data = counter_data,

  ci = 0.95

)

```

The resulting expected sum difference between the `baseline_data` and `counter_data`

are

```{r print result}

print(test_diffs)

```

The default setting is to estimate the confidence interval from the delta 

method. The resulting confidence interval from sampling of the posterior (with

improper priors) can be calculated as

```{r example 2}

post_diffs <- calc_sum_counterfactual_gam(m, 

  baseline_data,

  use_post = TRUE,

  nrep = 1000,

  counter_data = counter_data,

  ci = 0.95

)

print(post_diffs)

```

## Code of Conduct

  

  Please note that the gamdiffs project is released with a [Contributor Code of Conduct](https://contributor-covenant.org/version/2/0/CODE_OF_CONDUCT.html). By contributing to this project, you agree to abide by its terms.