BayesMFSurv

BayesMFSurv: An R Package to Estimate Bayesian Split-Population Survival Models With (and Without) Misclassified Failure Events - Published in JOSS (2020)

https://github.com/nicolas-schmidt/bayesmfsurv

Keywords

Repository

Bayesian Misclassified-Failure Survival Model

README.Rmd

---
output: github_document
---



```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
```

## The `BayesMFSurv` package

_Minnie M. Joo, Nicolas Schmidt, Sergio Bejar, Bumba Mukherjee, Vineeta Yadav_




[![CRAN\_Status\_Badge](https://www.r-pkg.org/badges/version/BayesMFSurv)](https://cran.r-project.org/package=BayesMFSurv)
`r badger::badge_devel("Nicolas-Schmidt/BayesMFSurv", "blue")`
[![R build status](https://github.com/Nicolas-Schmidt/BayesMFSurv/workflows/R-CMD-check/badge.svg)](https://github.com/Nicolas-Schmidt/BayesMFSurv/actions)
[![Project Status: Active – The project has reached a stable, usable state and is being actively developed.](https://www.repostatus.org/badges/latest/active.svg)](https://www.repostatus.org/#active)
[![Travis build status](https://travis-ci.com/Nicolas-Schmidt/BayesMFSurv.svg?branch=master)](https://travis-ci.com/Nicolas-Schmidt/BayesMFSurv)
[![CRAN RStudio mirrordownloads](https://cranlogs.r-pkg.org/badges/grand-total/BayesMFSurv?color=blue)](https://www.r-pkg.org/pkg/BayesMFSurv)
[![CRAN RStudio mirrordownloads](https://cranlogs.r-pkg.org/badges/BayesMFSurv?color=blue)](https://www.r-pkg.org/pkg/BayesMFSurv)
[![DOI](https://zenodo.org/badge/226979997.svg)](https://zenodo.org/badge/latestdoi/226979997)
[![DOI](https://joss.theoj.org/papers/10.21105/joss.02164/status.svg)](https://doi.org/10.21105/joss.02164)



### Description

Contains a split population survival estimator that models the misclassification probability of failure versus right-censored events. The split population survival estimator is described in Bagozzi et al. (2019) .

### Installation

You can install the released version (`0.1.0`) of `BayesMFSurv` from [CRAN](https://cran.r-project.org/) with:

``` r
install.packages("BayesMFSurv")
```

And the development version (`0.2.0`) from GitHub with:

``` r
if (!require("remotes")) install.packages("remotes")
remotes::install_github("Nicolas-Schmidt/BayesMFSurv")
```

## Functions


| Function |Description |
| ----------------- | -------------------------------------------------------------- | 
|`mfsurv`|fits a parametric Bayesian MF model via Markov Chain Monte Carlo (MCMC) to estimate the probability of misclassification in the first stage and the hazard in the second stage. Slice sampling is employed to draw the posterior sample of the model's split and survival stage parameters.|
|`mcmcsurv` | estimates a Bayesian Exponential or Weibull survival model via Markov Chain Monte Carlo (MCMC). Slice samplig is employed to draw the posterior sample of the model's survival stage parameters.|
|`stats`|a function to calculate the deviance information criterion (DIC) and the log-likelihood for fitted model objects of class mfsurv or mcmcsurv.|
|`summary`|returns a summary of a mfsurv or mcmcsurv object via `coda::summary.mcmc`.|



### Example


The data used to estimate the following examples come from Reenock, Bernhard and Sobek (2007) -DOI: 10.111/j.1468-2478.2007.00469.x-.  The RBS (2007) dataset uses continuous-time event history techniques to code episodes of democratic breakdown in all democracies from 1961 to 1995.  In addition, it provides data on a number of economic and political variables. 

|Variable| Description|
| ----------------- | -------------------------------------------------------------- | 
|__calinv__|inverse of per capita daily caloric supply|
|__lnlevel__|natural log of economic development|
|__calileve__|inverse of per capita daily caloric supply*lnlevel|
|__necon__|economic performance|
|__presi__|presidential regime|
|__tag__|effective number of parties|
|__rel__|religious fractionalization|
|__ethn__|ethnic fractionalization|
|__prevdem__|numbers of previous democratic episodes|
|__openc__|trade openness|
|__Y__|years in current democratic episode |
|__Y0__|years in current democratic episode (lagged)|
|__C__|breakdown of democratic episode|


#### Misclassified-Failure

`mfsurv` estimated the probability of misclassification failure in the first (split) stage and hazard in the second (survival) stage. 

`mfsurv` should be used when user suspects that data of survival cases could be right-sensored (i.e. when there is a probability that failure events are misclassified). 

Example with N = 100000 is [here](https://github.com/Nicolas-Schmidt/BayesMFSurv/tree/master/data-raw).

```{r}

# Baseline Bayesian misclassified failure (MF) model. 
# Misclassification stage only includes the intercept while the survival stage 
# includes all covariates described above.  

library(BayesMFSurv)

set.seed(95)
data(RBS)
RBS <- na.omit(RBS)
Y   <- RBS$Y
X   <- as.matrix(cbind(1, RBS[,1:10]))
C   <- RBS$C
Z1  <- cbind(rep(1,nrow(RBS)))
Y0  <- RBS$Y0
model1 <- mfsurv(Y ~ X | C ~ Z1, Y0 = Y0,
                 N = 100,
                 burn = 50,
                 thin = 5,
                 w = c(0.5, 0.5, 0.5),
                 m = 20,
                 form = 'Weibull')


stats(model1)

summary(model1, parameter = c("betas"))

```

#### Non Misclassified-Failure

`mcmcsurv` estimates a Bayesian equivalent of standard survival models (i.e. Exponential or Weibull).   

Example with N = 15000 can be found [here]( https://github.com/Nicolas-Schmidt/BayesMFSurv/tree/master/data-raw).

```{r}
set.seed(95)
model2 <- mcmcsurv(Y = Y, Y0 = Y0, C =  C,  X = X, 
                   N = 100, 
                   burn = 50, 
                   thin = 5, 
                   w = c(0.5, 0.5, 0.5),
                   m = 20, 
                   form = 'Weibull')


stats(model2)

summary(model2, parameter = c("betas"))
```




#### Citation

To cite package`BayesMFSurv` in publications, please use:

```{r}
citation(package = 'BayesMFSurv')
```

Owner

JOSS Publication

BayesMFSurv: An R Package to Estimate Bayesian Split-Population Survival Models With (and Without) Misclassified Failure Events

Published

April 24, 2020

DOI

10.21105/joss.02164

Volume 5, Issue 48, Page 2164

Authors

Minnie M. Joo
Dept. of Political Science, University of Massachusetts Lowell

Nicolás Schmidt
Dept. of Political Science. Universidad de la Republica, UY

Sergio Béjar
Dept. of Political Science, San Jose State University

Vineeta Yadav
Dept. of Political Science, Pennsylvania State University

Bumba Mukherjee
Dept. of Political Science, Pennsylvania State University

Editor

Marcos Vital

Tags

survival analysis bayesian statistics

View PDF Review Thread Software Archive

GitHub Events

Committers

Last synced: 7 months ago

Issues and Pull Requests

Last synced: 6 months ago