In ElliotWinters/exwProject3: STAT 302 Project 3

knitr::opts_chunk$set(echo = TRUE)
library(readr)
library(exwProject3)
library(dplyr)
library(ggplot2)

my_penguins <- read.csv("C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/data/my_penguins.csv")
my_penguins_class <- read.csv("C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/data/my_penguins_class.csv")
my_penguins_train <- read.csv("C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/data/my_penguins_train.csv")

source("C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/R/my_rf_cv.R")

data(my_penguins)
data(my_penguins_train)
data(my_penguins_class)

rf_vector <- vector()
for (k in c(2, 5, 10)){
  for (j in 1:30){
    rf_vector <- append(rf_vector, my_rf_cv(k))
  }
}

k_vector <- rep(2, 30)
k_vector <- append(k_vector, rep(5, 30))
k_vector <- append(k_vector, rep(10, 30))

rf_MSE <- data.frame("MSE" = rf_vector,
                     "k" = as.factor(k_vector))


plot_rf_cv <- ggplot(data = rf_MSE,
                     aes(x = k, y = MSE)) +
                geom_boxplot() +
                theme_bw(base_size = 16) + 
                labs(title = "MSE by number of folds (n=30)",
                     x = "k (num. folds)")

ggsave(path = "C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/Output/Figures",
       filename = "plot_rf_cv.png")

write.csv(rf_vector, "C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/Output/Results/simulation_results.csv")

mean_vec <- vector()
sd_vec <- vector()
for (m in c(2, 5, 10)){
  mean_vec <- append(mean_vec, mean(dplyr::filter(rf_MSE, k == m)[,1]))
  sd_vec <- append(sd_vec, sd(dplyr::filter(rf_MSE, k == m)[,1]))
}

rf_sd_table <- data.frame("k" = c(2, 5, 10),
                          "Mean" = mean_vec,
                          "SD" = sd_vec)

saveRDS(rf_sd_table, "C:/Users/Elliot/Desktop/STAT 302/projects/project_3/exwProject3/Output/Results/summary_table.rds")