R probably int_conformal_split 通過分割共形推理預測區間

int_conformal_split(object, ...)

# S3 method for default
int_conformal_split(object, ...)

# S3 method for workflow
int_conformal_split(object, cal_data, ...)

predict.int_conformal_split()

library(workflows)
library(dplyr)
library(parsnip)
library(rsample)
library(tune)
library(modeldata)

set.seed(2)
sim_train <- sim_regression(500)
sim_cal   <- sim_regression(200)
sim_new   <- sim_regression(  5) %>% select(-outcome)

# We'll use a neural network model
mlp_spec <-
  mlp(hidden_units = 5, penalty = 0.01) %>%
  set_mode("regression")

mlp_wflow <-
  workflow() %>%
  add_model(mlp_spec) %>%
  add_formula(outcome ~ .)

mlp_fit <- fit(mlp_wflow, data = sim_train)

mlp_int <- int_conformal_split(mlp_fit, sim_cal)
mlp_int
#> Split Conformal inference
#> preprocessor: formula 
#> model: mlp (engine = nnet) 
#> calibration set size: 200 
#> 
#> Use `predict(object, new_data, level)` to compute prediction intervals

predict(mlp_int, sim_new, level = 0.90)
#> # A tibble: 5 × 3
#>   .pred .pred_lower .pred_upper
#>   <dbl>       <dbl>       <dbl>
#> 1  4.46       -27.5        36.4
#> 2  5.83       -26.1        37.8
#> 3  9.27       -22.7        41.2
#> 4  1.50       -30.4        33.4
#> 5  9.68       -22.3        41.6