## ----setup, include = FALSE---------------------------------------------------
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

## ----load, warning=FALSE, message=FALSE---------------------------------------
library(NHSRdatasets)
library(dplyr)
library(ggplot2)
library(caret)
library(rsample)
library(varhandle)
library(tibble)

stranded_data <- NHSRdatasets::stranded_data

tibble::glimpse(stranded_data)

prop.table(table(stranded_data$stranded.label))

## ----feature_engineering------------------------------------------------------
stranded_data <- stranded_data |>
  dplyr::mutate(stranded.label = factor(stranded.label)) |>
  dplyr::select(everything(), -c(admit_date))

## ----dummy_encode-------------------------------------------------------------
cats <- select_if(stranded_data, is.character)

cat_dummy <- varhandle::to.dummy(cats$frailty_index, "frail_ind")

# Converts the frailty index column to dummy encoding and sets a column called "frail_ind" prefix
cat_dummy <- cat_dummy |>
  as.data.frame() |>
  dplyr::select(-frail_ind.No_index_item) # Drop the field of interest

# Drop the frailty index from the stranded data frame and bind on our new encoding categorical variables
stranded_data <- stranded_data |>
  dplyr::select(-frailty_index) |>
  bind_cols(cat_dummy) |>
  na.omit(.)

## ----train_test_split---------------------------------------------------------
split <- rsample::initial_split(stranded_data, prop = 3 / 4)
train <- rsample::training(split)
test <- rsample::testing(split)

## ----class_model--------------------------------------------------------------
# Fix the seed so that the same numbers will always be generated
set.seed(123)

glm_class_mod <- caret::train(factor(stranded.label) ~ .,
  data = train,
  method = "glm"
)


glm_class_mod

## ----predicting---------------------------------------------------------------
preds <- predict(glm_class_mod, newdata = test) # Predict class
pred_prob <- predict(glm_class_mod, newdata = test, type = "prob") # Predict probs

# Join prediction on to actual test data frame and evaluate in confusion matrix

predicted <- data.frame(preds, pred_prob)

test <- test |>
  bind_cols(predicted) |>
  dplyr::rename(pred_class = preds)

glimpse(test)

## ----evaluation---------------------------------------------------------------
caret::confusionMatrix(test$stranded.label, test$pred_class, positive = "Stranded")