R Markdown

—- Install and Load Necessary Packages —-

if (!requireNamespace(“ISLR2”, quietly = TRUE)) install.packages(“ISLR2”) if (!requireNamespace(“GGally”, quietly = TRUE)) install.packages(“GGally”) if (!requireNamespace(“FNN”, quietly = TRUE)) install.packages(“FNN”)

library(ISLR2) library(GGally) library(FNN)

—- Load Datasets —-

data(“Boston”) data(“Carseats”) data(“Auto”)

—- Chapter 2: Boston Housing Data —-

Dataset Summary

summary(Boston) dim(Boston) # Rows and Columns

Pairwise Scatterplots

ggpairs(Boston)

Crime Rate Analysis

cor(Boston$crim, Boston)

High Crime Rate, Tax Rate, Pupil-Teacher Ratio

summary(Boston\(crim) summary(Boston\)tax) summary(Boston$ptratio)

Census Tracts Adjacent to the Charles River

sum(Boston$chas == 1)

Median Pupil-Teacher Ratio

median(Boston$ptratio)

Census Tract with Lowest Median Home Value

Boston[which.min(Boston$medv), ]

Census Tracts with High Room Counts

sum(Boston\(rm > 7) sum(Boston\)rm > 8) Boston[Boston$rm > 8, ]

—- Chapter 3: KNN Classifier vs Regression —-

Sample Data for KNN Regression and Classification

train_data <- data.frame(x = c(1, 2, 3, 4, 5), y = c(10, 15, 20, 25, 30)) test_data <- data.frame(x = c(2.5, 3.5))

KNN Regression using FNN package

knn_regression <- knn.reg(train = data.frame(x = train_data\(x), test = data.frame(x = test_data\)x), y = train_data$y, k = 3)

print(knn_regression$pred) # Display predicted values

—- Chapter 3: Carseats Regression —-

Fit Multiple Regression Model

model_carseats <- lm(Sales ~ Price + Urban + US, data = Carseats) summary(model_carseats)

Confidence Intervals for Coefficients

confint(model_carseats)

Reduced Model Including Only Significant Predictors

model_refined <- lm(Sales ~ Price + US, data = Carseats) summary(model_refined)

Compare Model Performance

summary(model_carseats)\(adj.r.squared summary(model_refined)\)adj.r.squared

Outliers or High Leverage Observations

plot(model_refined)

—- Chapter 4: Logistic Regression vs Softmax —-

Define Variable x for Logistic Regression

x <- seq(-10, 10, length.out = 100)

Logistic Model Log-Odds

beta_0 <- 2 beta_1 <- -1 log_odds <- beta_0 + beta_1 * x

Softmax Model Coefficients

alpha_orange_0 <- beta_0 + 3 alpha_orange_1 <- beta_1 - 0.6 alpha_apple_0 <- 3 alpha_apple_1 <- 0.6

log_odds_softmax <- (alpha_orange_0 - alpha_apple_0) + (alpha_orange_1 - alpha_apple_1) * x