04b_combine_ROCs.Rmd

---
title: "Enet and PSO-SVM ROCs"
author: "Kat Moore"
date: "`r Sys.Date()`"
output: 
  html_document:
    toc: yes
    toc_float: yes
    toc_depth: 5
    df_print: paged
    highlight: kate
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)

library(here)
library(pROC)
library(tidyverse)

theme_set(theme_bw())
```

We've already shown ROCs for the enet and pso-svm individually; now let's show them together.

## Load data

```{r}
enet <- read_csv(here("04_enet_sample_predictions.csv"))

head(enet)
```

```{r}
pso <- read_csv(here("03_pso_sample_predictions.csv"))

head(pso)
```

## Combined ROCs

Remember:

>with direction = '>', observations are positive when they are smaller than or equal (<=) to the threshold. With direction = '<', observations are positive when they are greater than or equal (>=) to the threshold.

```{r}
roc_enet <- pROC::roc(response = enet$true.group,
                      levels = c("healthyControl", "breastCancer"),
                      predictor = enet$prob.breastCancer,
                      direction = "<")
roc_pso <- pROC::roc(response = pso$real.group,
                   levels = c("healthyControl", "breastCancer"),
                   predictor = pso$breastCancer,
                   direction = "<")

pROC::ggroc(list("elastic net" = roc_enet, "PSO-SVM" = roc_pso)) +
  ggsci::scale_color_lancet(name = "Classifiers") +
  ggtitle("Classifier performance on internal validation set") +
  #Add enet AUC using same color scale
  annotate(geom="text", x=0.25, y=0.5,
           label=paste("AUC:",signif(pROC::auc(roc_enet),3)),
           color=ggsci::pal_lancet()(2)[1]) +
  #Add pso AUC using same color scale
  annotate(geom="text", x=0.25, y=0.4,
           label=paste("AUC:",signif(pROC::auc(roc_pso),3)),
           color=ggsci::pal_lancet()(2)[2])
```

Without ggplot, it's easier to extract the AUCs, but color and positioning are more difficult to adjust.

```{r}
plot2rocs <- function(roc1, roc2, title){
  plot.roc(roc1,
           main = title, 
           percent=TRUE,
           #ci = TRUE,  #Finite xlim glitch 
           print.auc = TRUE,
           asp = NA)
  plot.roc(roc2, print.auc=T, add=T)
  
}

plot2rocs(roc1 = pROC::roc(response = enet$true.group,
                   levels = c("healthyControl", "breastCancer"),
                   predictor = enet$prob.breastCancer,
                   direction = "<"),
          roc2 = pROC::roc(response = pso$real.group,
                   levels = c("healthyControl", "breastCancer"),
                   predictor = pso$breastCancer,
                   direction = "<"),
          title="Classifier performance on internal validation set")
```

```{r}
sessionInfo()
```