You can install the development version of {bicopula} from GitHub with:
# install.packages("devtools")
devtools::install_github("ellessenne/bicopula")This is a basic example. Simulating some bivariate data:
library(bicopula)
#> Loading required package: survival
set.seed(42)
N <- 500
df <- data.frame(
age = runif(N, 20, 40),
sex = rbinom(N, size = 1, prob = 0.5),
biomarker = runif(N, 0, 120)
)
simdata <- simulate_bisurv(
dist1 = Exponential$new(lambda = 0.1),
dist2 = Weibull$new(lambda = 0.1, gamma = 1.5),
formula1 = ~ age + sex,
formula2 = ~ age + biomarker,
data = df,
beta1 = c(age = 0.01, sex = 0.5),
beta2 = c(age = 0.01, biomarker = 0.005),
copula = copula::frankCopula(param = 5)
)
head(simdata)
#> eventtime1 status1 eventtime2 status2
#> 1 19.9125778 1 8.4047075 1
#> 2 0.3118698 1 0.9653524 1
#> 3 17.1881459 1 5.2461692 1
#> 4 0.3472099 1 2.2786107 1
#> 5 7.2249324 1 4.5745431 1
#> 6 1.6068254 1 2.9668163 1
with(simdata, plot(eventtime1, eventtime2))
Then, fitting a bivariate survival model:
df <- cbind(df, simdata)
bisurvreg(
formula1 = Surv(eventtime1, status1) ~ age + sex,
distribution1 = "exponential",
formula2 = Surv(eventtime2, status2) ~ age + biomarker,
distribution2 = "weibull",
copula = "frank",
data = df
)
#> $par
#> (Intercept) age sex (Intercept) age biomarker
#> -2.933330709 0.009289732 0.360414799 -3.047869671 0.008824681 0.004833773
#> theta ancillary
#> 4.110922938 0.448150921
#>
#> $objective
#> [1] 2097.916
#>
#> $convergence
#> [1] 1
#>
#> $iterations
#> [1] 5
#>
#> $evaluations
#> function gradient
#> 6 5
#>
#> $message
#> [1] "singular convergence (7)"Lots of work needs to be done to make this pretty, but as a proof-of-concept it doesn’t feel too bad!