Scribd
Upload
12 views

R-code

It is about coding in r for robus regression analysis for our last semestee courae

Uploaded by

Fariya Ahmed
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
12 views

R-code

It is about coding in r for robus regression analysis for our last semestee courae

Uploaded by

Fariya Ahmed
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 5

R-code for Question 1

# Load necessary libraries


library(MASS)
library(car)

# Generate Dataset
set.seed(123)
n <- 50
x1 <- rnorm(n, mean=3, sd=0.75)
x2 <- rnorm(n, mean=5, sd=0.5)
y <- 3 + 2 * x1 - 1.5 * x2 + rnorm(n, mean=0, sd = 3)
data <- data.frame(y, x1, x2)

# Fit OLS model

ols_model <- lm(y ~ x1 + x2, data = data)


summary(ols_model)

# Fit robust regression model

robust_model <- rlm(y ~ x1 + x2, data = data, method = "M")


summary(robust_model)

# Residual Analysis

residuals <- residuals(robust_model)

# Plot residuals to check for patterns

par(mfrow = c(2, 2))


plot(residuals, main = "Robust Residuals", ylab = "Residuals", xlab = "Index")
abline(h = 0, col = "red", lty = 2)

# Plot to check for normality of residuals

hist(residuals, main = "Residuals Histogram", xlab = "Residuals", breaks = 10)

qqnorm(residuals, main = "Q-Q Plot of Residuals")


qqline(residuals, col = "red")

# Plot Residuals vs Fitted Values (Homoscedasticity check)

plot(fitted(robust_model), residuals, main = "Residuals vs Fitted", xlab =


"Fitted Values", ylab = "Residuals")
abline(h = 0, col = "red", lty = 2)

# Standardized residuals (for detecting outliers) (Residuals > 2 are flagged)

robust_std_res <- residuals/summary(robust_model)$sigma


outliers <- which(abs(robust_std_res) > 2)
cat(outliers)

# Highlight outliers in residual plot

plot(fitted(robust_model), residuals, main = "Outliers Highlighted", xlab =


"Fitted Values", ylab = "Residuals", col = ifelse(abs(robust_std_res) > 2, "red",
"black"))
abline(h = 0, col = "blue", lty = 2)

# Influence Diagnostics
# Leverage: Hat values

leverage <- hatvalues(robust_model)


plot(leverage, main = "Leverage Plot", ylab = "Leverage", xlab = "Index", col =
ifelse(leverage > 2 * mean(leverage), "red", "black"))
abline(h = 2 * mean(leverage), col = "red", lty = 2)

#Identify leverage ponits (# Threshold for leverage (2 * mean leverage))

leverage_points <- which(leverage > 2 * mean(leverage))


cat(leverage_points)

# Cook's Distance

cooks_d <- cooks.distance(ols_model)


plot(cooks_d, main = "Cook's Distance", ylab = "Cook's Distance", xlab = "Index",
col = ifelse(cooks_d > 4/n, "red", "black"))
abline(h = 4 / n, col = "red", lty = 2)

# Identify influential points

influential_points <- which(cooks_d > (4 / n))


cat(influential_points)

R code for Question 2:

library(MASS)
library(car)

data(mtcars)
head(mtcars)

#Fit a linear model

ols_model <- lm(mpg ~ wt + hp, data = mtcars)

summary(ols_model)

# Fit a robust regression model

robust_model <- rlm(mpg ~ wt + hp, data = mtcars, method="M")

summary(robust_model)

# Residual analysis

residuals <- residuals(robust_model)

# Standardized residuals (for detecting outliers)

std_residuals <- residuals/summary(robust_model)$sigma


outliers <- which(abs(std_residuals) > 2)
cat(outliers)

# Highlight outliers in residual plot

par(mfrow = c(2, 2))


plot(fitted(robust_model), residuals, main = "Outliers Highlighted", xlab =
"Fitted Values", ylab = "Residuals", col = ifelse(abs(std_residuals) > 2, "red",
"black"))
abline(h = 0, col = "blue", lty = 2)

# Leverage values

leverage <- hatvalues(robust_model)


leverage_points <- which(leverage > 2 * mean(leverage))
cat(leverage_points)

plot(leverage, main = "Leverage Plot", ylab = "Leverage", xlab = "Index", col =


ifelse(leverage > 2 * mean(leverage), "red", "black"))
abline(h = 2 * mean(leverage), col = "red", lty = 2)

# Cook's distance

cooks_d <- cooks.distance(ols_model)


influential_points <- which(cooks_d > (4 / length(mtcars$mpg)))
cat(influential_points)

plot(cooks_d, main = "Cook's Distance", ylab = "Cook's Distance", xlab = "Index",


col = ifelse(cooks_d > 4/length(mtcars$mpg), "red", "black"))
abline(h = 4 / length(mtcars$mpg), col = "red", lty = 2)

R-code for Question 3:

df <- data.frame(
x1 = c(1, 3, 3, 4, 4, 6, 6, 8, 9, 3, 11, 16, 16, 18, 19, 20, 23, 23, 24, 25),
x2 = c(7, 7, 4, 29, 13, 34, 17, 19, 20, 12, 25, 26, 26, 26, 27, 29, 30, 31, 31,
32),
y = c(17, 170, 19, 194, 24, 2, 25, 29, 30, 32, 44, 60, 61, 63, 63, 64, 61, 67,
59, 70)
)

# Fit the linear model

ols_model <- lm(y ~ x1 + x2, data = df)

summary(ols_model)

# Load the MASS package

library(MASS)

# Fit the robust regression model


robust_model <- rlm(y ~ x1 + x2, data = df)

summary(robust_model)

#Compare between lm & rlm by residual standard error

summary(ols_model)$sigma
summary(robust_model)$sigma

# Residual analysis

residuals <- residuals(robust_model)

# Standardized residuals (for detecting outliers)

std_residuals <- residuals/summary(robust_model)$sigma


outliers <- which(abs(std_residuals) > 2)
cat(outliers)

# Highlight outliers in residual plot

par(mfrow = c(2, 2))


plot(fitted(robust_model), residuals, main = "Outliers Highlighted", xlab =
"Fitted Values", ylab = "Residuals", col = ifelse(abs(std_residuals) > 2, "red",
"black"))
abline(h = 0, col = "blue", lty = 2)

# Leverage values

leverage <- hatvalues(robust_model)


leverage_points <- which(leverage > 2 * mean(leverage))
cat(leverage_points)

plot(leverage, main = "Leverage Plot", ylab = "Leverage", xlab = "Index", col =


ifelse(leverage > 2 * mean(leverage), "red", "black"))
abline(h = 2 * mean(leverage), col = "red", lty = 2)

# Cook's distance

cooks_d <- cooks.distance(ols_model)


influential_points <- which(cooks_d > (4 / length(df$y)))
cat(influential_points)

plot(cooks_d, main = "Cook's Distance", ylab = "Cook's Distance", xlab = "Index",


col = ifelse(cooks_d > 4/length(df$y), "red", "black"))
abline(h = 4 / length(df$y), col = "red", lty = 2)

R-code for Question 4:

# Install and load the quantreg package


install.packages("quantreg")
library(quantreg)
# Data
data(mtcars)

# Perform quantile regression at the 0.25th, 0.5th (median), and 0.75th quantiles
qr_model25 <- rq(mpg ~ wt + hp, data = mtcars, tau = 0.25)
qr_model50 <- rq(mpg ~ wt + hp, data = mtcars, tau = 0.50)
qr_model75 <- rq(mpg ~ wt + hp, data = mtcars, tau = 0.75)

# Display results
summary(qr_model25)
summary(qr_model50)
summary(qr_model75)

# Fit quantile regression for multiple quantiles

qr_multiple <- rq(mpg ~ wt + hp, data = mtcars, tau = c(0.25, 0.50, 0.75))
summary(qr_multiple)

You might also like