Module_8_Lab_Tewary_Raushan

Load Necessary Libraries

library(tidymodels)
library(ggplot2)
library(readr)
library(dplyr)

# Ensure 'kknn' is installed for KNN models
if (!requireNamespace("kknn", quietly = TRUE)) {
  install.packages("kknn")
}

## Load the Dataset
boston <- readr::read_csv("C:\\Users\\raush\\boston.csv", show_col_types = FALSE)

## Split the Data into Training and Test Sets
set.seed(123)
split <- initial_split(boston, prop = 0.7, strata = cmedv)
train <- training(split)
test <- testing(split)

# Print the number of observations
n_train <- nrow(train)
n_test <- nrow(test)

n_train

## [1] 352

n_test

## [1] 154

## Compare Distributions of cmedv
ggplot(train, aes(x = cmedv)) + 
  geom_density(color = "blue", fill = "blue", alpha = 0.3) + 
  geom_density(data = test, aes(x = cmedv), color = "red", fill = "red", alpha = 0.3) + 
  labs(title = "Distribution of cmedv in Training vs Test Set", x = "cmedv", y = "Density")

## Fit Linear Regression Models
# Model using 'rm' to predict 'cmedv'
lm1 <- linear_reg() %>%
  set_engine("lm") %>%
  fit(cmedv ~ rm, data = train)

# Compute RMSE for Model 1
lm1 %>%
  predict(test) %>%
  bind_cols(test %>% select(cmedv)) %>%
  rmse(truth = cmedv, estimate = .pred)

## # A tibble: 1 × 3
##   .metric .estimator .estimate
##   <chr>   <chr>          <dbl>
## 1 rmse    standard        6.83

# Model using all features to predict 'cmedv'
lm2 <- linear_reg() %>%
  set_engine("lm") %>%
  fit(cmedv ~ ., data = train)

# Compute RMSE for Model 2
lm2 %>%
  predict(test) %>%
  bind_cols(test %>% select(cmedv)) %>%
  rmse(truth = cmedv, estimate = .pred)

## # A tibble: 1 × 3
##   .metric .estimator .estimate
##   <chr>   <chr>          <dbl>
## 1 rmse    standard        4.83

## Fit a KNN Model
knn <- nearest_neighbor() %>%
  set_engine("kknn") %>%
  set_mode("regression") %>%
  fit(cmedv ~ ., data = train)

# Compute RMSE for the KNN model
knn %>%
  predict(test) %>%
  bind_cols(test %>% select(cmedv)) %>%
  rmse(truth = cmedv, estimate = .pred)

## # A tibble: 1 × 3
##   .metric .estimator .estimate
##   <chr>   <chr>          <dbl>
## 1 rmse    standard        3.37

## Summary Statistics and Missing Values
# Check for missing values
missing_values <- sum(is.na(boston$cmedv))
missing_values

## [1] 0

# Calculate minimum, maximum, and average cmedv
summary_stats <- boston %>%
  summarise(
    minimum_cmedv = min(cmedv, na.rm = TRUE),
    maximum_cmedv = max(cmedv, na.rm = TRUE),
    average_cmedv = mean(cmedv, na.rm = TRUE)
  )
summary_stats

## # A tibble: 1 × 3
##   minimum_cmedv maximum_cmedv average_cmedv
##           <dbl>         <dbl>         <dbl>
## 1             5            50          22.5

# Calculate the median of cmedv
median_cmedv <- median(boston$cmedv, na.rm = TRUE)
print(median_cmedv)

## [1] 21.2