boot_predict: Bootstrap simulation for model prediction

In finalfit: Quickly Create Elegant Regression Results Tables and Plots when Modelling

Bootstrap simulation for model prediction

Description

Generate model predictions against a specified set of explanatory levels with bootstrapped confidence intervals. Add a comparison by difference or ratio of the first row of newdata with all subsequent rows.

Usage

boot_predict(
  fit,
  newdata,
  type = "response",
  R = 100,
  estimate_name = NULL,
  confint_level = 0.95,
  conf.method = "perc",
  confint_sep = " to ",
  condense = TRUE,
  boot_compare = TRUE,
  compare_name = NULL,
  comparison = "difference",
  ref_symbol = "-",
  digits = c(2, 3)
)

Arguments

fit	A model generated using lm, glm, lmmulti, and glmmulti.
newdata	Dataframe usually generated with finalfit_newdata.
type	the type of prediction required, see predict.glm. The default for glm models is on the scale of the response variable. Thus for a binomial model the default predictions are predicted probabilities.
R	Number of simulations. Note default R=100 is very low.
estimate_name	Name to be given to prediction variable y-hat.
confint_level	The confidence level to use for the confidence interval. Must be strictly greater than 0 and less than 1. Defaults to 0.95, which corresponds to a 95 percent confidence interval.
conf.method	Passed to the type argument of boot::boot.ci(). Defaults to "perc". The allowed types are "perc", "basic", "bca", and "norm". Does not support "stud" or "all"
confint_sep	String separating lower and upper confidence interval
condense	Logical. FALSE gives numeric values, usually for plotting. TRUE gives table for final output.
boot_compare	Include a comparison with the first row of newdata with all subsequent rows. See boot_compare.
compare_name	Name to be given to comparison metric.
comparison	Either "difference" or "ratio".
ref_symbol	Reference level symbol
digits	Rounding for estimate values and p-values, default c(2,3).

Details

To use this, first generate newdata for specified levels of explanatory variables using finalfit_newdata. Pass model objects from lm, glm, lmmulti, and glmmulti. The comparison metrics are made on individual bootstrap samples distribution returned as a mean with confidence intervals. A p-value is generated on the proportion of values on the other side of the null from the mean, e.g. for a ratio greater than 1.0, p is the number of bootstrapped predictions under 1.0, multiplied by two so is two-sided.

Value

A dataframe of predicted values and confidence intervals, with the option of including a comparison of difference between first row and all subsequent rows of newdata.

See Also

finalfit_newdata

Examples

library(finalfit)
library(dplyr)

# Predict probability of death across combinations of factor levels
explanatory = c("age.factor", "extent.factor", "perfor.factor")
dependent = 'mort_5yr'

# Generate combination of factor levels
colon_s %>%
  finalfit_newdata(explanatory = explanatory, newdata = list(
    c("<40 years",  "Submucosa", "No"),
    c("<40 years", "Submucosa", "Yes"),
    c("<40 years", "Adjacent structures", "No"),
    c("<40 years", "Adjacent structures", "Yes")
   )) -> newdata

# Run simulation
colon_s %>%
  glmmulti(dependent, explanatory) %>%
  boot_predict(newdata, estimate_name = "Predicted probability of death",
    compare_name = "Absolute risk difference", R=100, digits = c(2,3))

# Plotting
explanatory = c("nodes", "extent.factor", "perfor.factor")
colon_s %>%
  finalfit_newdata(explanatory = explanatory, rowwise = FALSE, newdata = list(
  rep(seq(0, 30), 4),
  c(rep("Muscle", 62), rep("Adjacent structures", 62)),
  c(rep("No", 31), rep("Yes", 31), rep("No", 31), rep("Yes", 31))
)) -> newdata

colon_s %>%
  glmmulti(dependent, explanatory) %>%
  boot_predict(newdata, boot_compare = FALSE, R=100, condense=FALSE) -> plot

  library(ggplot2)
  theme_set(theme_bw())
  plot %>%
    ggplot(aes(x = nodes, y = estimate, ymin = estimate_conf.low,
        ymax = estimate_conf.high, fill=extent.factor))+
      geom_line(aes(colour = extent.factor))+
      geom_ribbon(alpha=0.1)+
      facet_grid(.~perfor.factor)+
      xlab("Number of postive lymph nodes")+
      ylab("Probability of death")+
      labs(fill = "Extent of tumour", colour = "Extent of tumour")+
      ggtitle("Probability of death by lymph node count")

finalfit documentation built on Sept. 11, 2024, 9:01 p.m.