ex5_2_4.R

# exercise 5.2.4

# Load wine data 
source(file.path("Scripts", "ex5_1_5.R"))

# Fit linear regression model to predict Alcohol from all other attributes
y = X[, 11]
Xr = data.frame(X[, 1:10])
xnam <- paste("X", 1:10, sep="")
(fmla <- as.formula(paste("y ~ ", paste(xnam, collapse= "+"))))
w_est = lm(fmla, data=Xr)

# Make a scatter plot of predicted versus true values of Alcohol
y_est = w_est$fitted.values
plot(y, y_est, xlab='Alcohol (true)', ylab='Alcohol (estimated)', main='Alcohol content', pch=20)

# Make a histogram of the residual error
hist(y-y_est, breaks=41, main="Residual error");