Simulate from a logistic regression model

Usage

lr_simulator(object)

Arguments

object: A logistic regression model

Value

A function with arguments param, data, and type.

The param argument should be a vector of coefficients
The data argument should be a data frame or tibble
The type argument should be a string indicating the type of prediction to generate (defaults to "response")

Takes a fitted glm object as input and returns a function that will evaluate the underlying structural model with user-specified parameters or data (e.g., for VPCs or other counterfactual simulation scenarios). In principle this should work for glms more generally, not merely logistic regressions, but has not been tested except for logistic regression models

Examples

mod1 <- lr_model(response ~ exposure + sex, lr_data)
par1 <- coef(mod1)
mod1_sim <- lr_simulator(mod1)

# no counterfactuals
p1 <- mod1_sim(param = par1, data = lr_data) 
p2 <- unname(predict(mod1, type = "response")) # same result

# user modifies the data set
lr_data2 <- lr_data[1:20, ]
p3 <- mod1_sim(param = par1, data = lr_data2) 
p4 <- unname(predict(mod1, newdata = lr_data2, type = "response")) # same result

# user modifies the parameters
par2 <- par1
int1 <- par1["(Intercept)"]
par2["(Intercept)"] <- 0
p5 <- mod1_sim(param = par2, data = lr_data)